Gender Differences in Depression in Representative National Samples: Meta-Analyses of Diagnoses and Symptoms

Classified as probably not a substantial claim.

A study including 18 countries from the WHO World Mental Health Surveys reported that the relationship between gender and major depressive episode did not differ significantly between high-income and low- to middle-income countries.

“A study including 18 countries from the WHO World Mental Health Surveys (Kessler & Usten, 2008) reported that the relationship between gender and MDE did not differ significantly between high-income and low- to middle-income countries”

• cites Bromet et al., 2011 sourceSupported

May combine two claims.

The 2-to-1 ratio for major depression is not universal and may vary substantially across nations.

“However, findings also suggest that the 2:1 ratio is not universal and may vary substantially across nations.”

• cites Bromet et al., 2011 sourceSupported

May combine two claims.

In most adult epidemiological reports, women have higher rates of major depression than men, with an average ratio of 2 to 1.

“In the vast majority of epidemiological reports on adults, women have higher rates of major depression compared to men; on average, the ratio is 2:1”

• cites Andrade et al., 2003 sourceNot checked — no full text of the source
• cites Bromet et al., 2011 sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing).

Other widely held beliefs about gender differences, such as the gender difference in math performance, have sometimes been found inaccurate when meta-analyzed.

“Other widely held beliefs about gender differences, such as the gender difference in math performance, have sometimes been found to be inaccurate when the data are meta-analyzed”

• cites Hyde et al., 2008Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Lindberg et al., 2010 sourceSupported

Classified as probably not a substantial claim. May combine two claims.

There is increased recognition of the validity and value of dimensional assessments, as well as the impairment associated with sub-threshold symptoms that do not meet diagnostic criteria.

“there is increased recognition of the validity (e.g., Markon, Chmielewski, & Miller, 2011) and value (Cuthbert & Insel, 2013) of dimensional assessments, as well as the impairment associated with sub-threshold levels of symptoms that do not meet diagnostic criteria.”

• cites Markon et al., 2011 sourceNot checked — no full text of the source
• cites Cuthbert & Insel, 2013 sourceSupported

May combine two claims.

Two large multi-nation studies reported conflicting results, finding that the gender gap in depression was smaller and larger, respectively, in low gender-equity countries compared to high gender-equity countries.

“Two large multi-nation studies have reported conflicting results, finding that the gender gap in depression was smaller and larger, respectively, in low gender-equity countries compared to high gender-equity countries”

• cites Hopcroft & Bradley, 2007 sourceNot checked — no full text of the source
• cites Van de Velde et al., 2013 sourcePartly supported

  The Discussion reports the conflicting finding that Hopcroft and Bradley found a larger depression gender gap in high-equality countries, while this paper's results were more consistent with equality reducing women's disadvantage. But the Results say the overall interaction between gender and equality was not significant; a smaller gap at higher equality appeared only in certain subgroups and for certain measures. The sentence overstates this paper's finding as a general cross-country result. (LLM, conf=0.87)

  Experimental — accuracy not yet measured

Classified as probably not a substantial claim.

Seedat and colleagues found a significant narrowing in the gender difference in depression in recent cohorts.

“Seedat and colleagues (2009) found a significant narrowing in the gender difference in depression in recent cohorts.”

• cites Seedat et al., 2009 sourceSupported

May combine two claims.

A latent growth curve analysis found that girls' depression symptoms accelerated early in adolescence and then leveled off, while boys' symptoms accelerated in late adolescence.

“One latent growth curve analysis indicated that girls’ depression symptoms accelerated early in adolescence and then leveled off, whereas boys’ symptom levels accelerated in late adolescence (Salk, Petersen, Abramson, & Hyde, 2016), consistent with the first possibility above.”

• cites Salk, Petersen, Abramson, & Hyde, 2016 sourceSupported

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Greater contraceptive prevalence should indicate women's greater reproductive control and a lower rate of unwanted pregnancies, which are a risk factor for depression.

“On the face of it, greater contraceptive prevalence should indicate women’s greater control over reproduction and a lower rate of unwanted pregnancies, which are a risk factor for depression (e.g., Mercier, Garrett, Thorp, & Siega-Riz, 2013).”

• cites Mercier, Garrett, Thorp, & Siega-Riz, 2013 sourcePartly supported

  The Discussion (Main Findings) reports that women with unwanted pregnancies had a higher risk of postpartum depression than women with intended pregnancies. The claim says “depression” without limiting it to the postpartum period; the paper also cautions in the Discussion (Interpretation) that it cannot establish a causal relationship. The contraceptive-prevalence premise appears to be the citing authors’ framing, not a finding attributed to this source. (LLM, conf=0.82)

  Experimental — accuracy not yet measured

Borderline on S3 (load-bearing), S4 (based): unclear from the text.

In both meta-analyses, U.S. ethnicity did not account for significant variation in effect size.

“In both meta-analyses, U.S. ethnicity did not account for significant variation in effect size (see Table 4).”

May combine two claims.

The meta-analysis of depression symptoms pooled 95 articles and their national datasets, covering 1,922,064 individuals from more than 90 countries.

“we meta-analyzed data from 65 and 95 articles and their corresponding national datasets, representing data from 1,716,195 and 1,922,064 people in over 90 different nations.”

May combine two claims.

For both meta-analyses, the gender difference declined after adolescence and remained stable in adulthood.

“but then declined and remained stable in adulthood.”

May combine two claims.

The meta-analysis of major depression diagnoses pooled 65 articles and their national datasets, covering 1,716,195 individuals from more than 90 countries.

“we meta-analyzed data from 65 and 95 articles and their corresponding national datasets, representing data from 1,716,195 and 1,922,064 people in over 90 different nations.”

May combine two claims.

The gender difference for major depression diagnoses emerged earlier than previously thought, with an odds ratio of 2.37 at age 12.

“The gender difference for diagnoses emerged earlier than previously thought, with OR = 2.37 at age 12.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Overall, 112 articles from the original search met inclusion criteria, and 46 new articles were added from additional searching for nationally representative datasets.

“Overall, 112 articles from the original search met criteria for inclusion, including articles for which authors needed to be contacted for data. We added 46 new articles that were not in the original search from additional searching for nationally representative datasets.”

May combine two claims.

Effect sizes for ages 8-12 (0.09), 13-15 (0.35), and 16-19 (0.41) were statistically different (QB = 23.01 and QB = 8.06, ps < .01), with each consecutive age group significantly larger than the previous one.

“Follow-up testing with pairs of consecutive age groups indicated that the effect sizes for ages 8–12 (0.09), 13–15 (0.35), and 16–19 (0.41) were statistically different (QB = 23.01 and QB = 8.06, ps< .01, respectively), such that the effect size for each consecutively age group was significantly larger than the previous age group.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The depression symptom meta-analysis covers data collected from 1978 to 2014, while the major depression meta-analysis covers data collected from 1991 to 2014.

“The depression symptom meta-analysis covers data collected from 1978–2014 whereas the major depression meta-analysis only includes data collected from 1991–2014.”

May combine two claims.

For major depression, effect sizes ranged from OR = 1.71 to OR = 3.02, with ORs above 2.0 during adolescence and between 1.71 and 2.02 in adulthood.

“Effect sizes ranged from OR = 1.71 to OR = 3.02, with ORs >2.0 during adolescence and ORs between 1.71 and 2.02 in adulthood.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In the major depression meta-analysis, 96% of articles reported DSM major depression diagnoses or episodes and 4% reported the roughly equivalent ICD-10 depressive episode.

“For the major depression meta-analysis, all articles reported DSM major depression diagnoses or episodes (96%) or the roughly equivalent ICD-10 depressive episode (4%).”

Probably not a substantial claim: may not pass S3 (load-bearing).

The random effects variance component for the major depression meta-analysis was 0.07.

“The random effects variance component was 0.07.”

May combine two claims.

The youngest age group available for the major depression analyses was 12 years old, so the emergence of the gender difference from childhood to adolescence could not be observed.

“Note that the youngest age group available for these analyses was 12 years old, making it impossible to observe the emergence of the gender difference from childhood to adolescence.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

When outliers were removed, income category became marginally significant (QB = 3.09, p = 0.08), with smaller gender differences in high-income nations (d = 0.26) than in low- to middle-income nations (d = 0.29).

“However, when outliers were removed, the effect became marginally significant (QB = 3.09, p = 0.08) with smaller gender differences in high-income nations (d = 0.26) compared to low- to middle- income nations (d = 0.29).”

Probably not a substantial claim: may not pass S3 (load-bearing).

The random effects variance component for the depression symptom meta-analysis was 0.02.

“The random effects variance component was 0.02.”

May combine two claims.

Although the likelihood ratio test showed a marginal effect for the major depression meta-analysis, the weighted averages from the unadjusted model (logged OR = 0.67) and adjusted model (logged OR = 0.62) were quite similar.

“Although the LRT showed a marginal effect for the major depression meta-analysis, the weighted average from the unadjusted model (logged OR = 0.67) and the adjusted model (logged OR = 0.62) were quite similar.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The symptom studies were published between 1991 and 2016, with data collected from 1978 to 2014 in 53 countries.

“These studies were published from 1991 to 2016 and yielded data from 1978–2014 collected from 53 different countries.”

May combine two claims.

The 95 symptom articles provided data comprising 180 samples, 413 effect sizes, and 1,922,064 people, of whom 52% were female.

“utilized data comprising 180 samples, 413 effect sizes, and 1,922,064 people (52% female).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Both meta-analyses include mostly high-income countries and cover the lifespan.

“They both include mostly high-income countries and cover the lifespan.”

May combine two claims.

The OR at ages 16-19 (2.69) was significantly larger than the OR at ages 20-29 (1.93), QB = 43.19, p < .001, indicating a significant decrease in the gender difference from adolescence to the 20s.

“The OR at ages 16–19 (2.69) was significantly larger than the OR for ages 20–29 (1.93), QB = 43.19, p < .001, indicating a significant decrease in the gender difference from adolescence to the 20s.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Removing outliers did not change the effect size estimates for depression symptoms.

“Removal of outliers did not change the estimates of effect sizes.”

May combine two claims.

Likelihood ratio tests comparing unadjusted and adjusted models for the major depression and depression symptom meta-analyses were not statistically significant (p = 0.073 and p = 0.3226, respectively).

“The likelihood ratio tests (LRT) comparing the unadjusted to adjusted models (using p-value cut points of 0.05, 0.01, and 0.001) for the major depression and depression symptom meta-analyses, respectively, were not statistically significant, p = 0.073 and p = 0.3226.”

May combine two claims.

The literacy variable was not range restricted, with ratios from .31 to 1.07.

“Importantly, there was not a range restriction for the literacy variable (ratios ranged from .31 to 1.07), which can have substantial negative skew (Else-Quest & Grabe, 2012).”

• cites Else-Quest & Grabe, 2012 sourceNot checked — no full text of the source

May combine two claims.

For major depression, contraceptive prevalence and literacy ratio both predicted variability in effect size.

“Contraceptive prevalence and literacy ratio both predicted variability in effect size.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The two meta-analyses differ in the distribution of nations and years of data collection.

“However, they differ in terms of the distribution of nations and year of data collection.”

Probably not a substantial claim: may not pass S3 (load-bearing).

In the symptom meta-analysis, most effect sizes were based on the CES-D measure.

“For the symptom meta-analysis, most effect sizes were based on the C ES-D measure.”

May combine two claims.

In follow-up analyses, effect sizes from gender-focused articles did not differ significantly from those focused on depression (QB = 2.22, p = 0.14) or on something other than gender or depression (QB = 0.91, p = 0.34).

“However, in follow-up analyses, effect sizes from articles that focused on gender were not significantly different than effect sizes from articles that focused on depression (QB = 2.22, p = 0.14) or something other than gender or depression (QB = 0.91, p =0.34).”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The diagnosis studies were published between 1993 and 2016, with data collected from 1991 to 2014 in 75 countries.

“These studies were published from 1993 to 2016 with data from 1991–2014 collected in 75 different countries.”

May combine two claims.

The 65 diagnosis articles provided data comprising 149 samples, 300 effect sizes, and 1,716,195 people, of whom 53% were female.

“provided data comprising 149 samples (this number includes different countries from the same international dataset as well as different years of data collection for cross-sectional national datasets), 300 effect sizes (this number includes separate effect sizes for different age groups and ethnicities), and 1,716,195 people (53% female).”

Probably not a substantial claim: may not pass S3 (load-bearing).

Most effect sizes in the major depression meta-analysis were 12-month major depressive episodes based on DSM-IV or DSM-IV-TR using a version of the CIDI.

“The vast majority of effect sizes for the major depression meta-analysis were 12-month major depressive episodes based on the DSM-IV or DSM-IV-TR using a version of the CIDI.”

May combine two claims.

When outliers were excluded, the OR for ages 13-15 decreased from 3.02 to 2.92 and the OR for age 70+ increased from 2.02 to 2.20.

“When outliers were excluded, the age 13–15 OR decreased from 3.02 to 2.92 and the age 70+ OR increased from 2.02 to 2.20.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The depression symptom meta-analysis includes the most effect sizes from Europe (53%).

“and the depression symptom meta-analysis includes the most effect sizes from Europe (53%).”

May combine two claims.

For depression symptoms, effect sizes ranged from d = 0.09 to d = 0.41, peaked at ages 16-19, declined in the 20s, and remained relatively stable at about d = 0.20 thereafter.

“Effect sizes ranged from d = 0.09 to d = 0.41, peaking at ages 16–19, declining in the 20s, and staying relatively stable at roughly d = 0.20 after that.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The diagnosis meta-analysis included 65 articles and their datasets, and the symptom meta-analysis included 95 articles and their datasets.

“The final sample of studies (see Figure 1) for the meta-analyses included data from 65 (diagnosis meta-analysis) and 95 (symptom meta-analysis) articles and their corresponding data sets.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

For depression symptoms, income inequality was a significant predictor, with larger gender differences reported in nations with low income inequality.

“Income inequality was a significant predictor of effect size, such that larger gender differences were reported in nations with low levels of income inequality.”

May combine two claims.

In the major depression meta-analysis, diagnostic manual, type of depression (episode versus disorder), and depression time span did not predict effect size variation.

“As shown in Table 4, manual for major depression (e.g., DSM-IV, ICD-10), type of depression (episode versus disorder) and depression time span (1 month, 6 months, 12 months, and lifetime) did not predict effect size variation.”

May combine two claims.

In both meta-analyses, the gender difference in depression peaked during adolescence, then declined and remained stable in adulthood.

“For both meta-analyses, the effect size peaked in adolescence but then declined and remained stable in adulthood, a finding that has not been identified previously.”

• Also stated in: § Conclusions · boosted

  “Our results provide powerful evidence that the gender difference in major depression diagnoses and depression symptoms peaks in adolescence, with the gender gap in diagnoses emerging earlier than previously thought (OR = 2.37 at age 12). The gender gap then narrows and remains stable in adulthood, a finding that has not been identified previously and has important implications for both theory and”

May combine two claims.

In both meta-analyses, the gender difference declined in early adulthood and then remained relatively stable, with odds ratios between 1.71 and 2.02 and d values between 0.19 and 0.30.

“In both meta-analyses, the gender difference declined in early adulthood and then remained relatively stable, hovering between OR = 1.71 – 2.02 and d = 0.19 – 0.30.”

May combine two claims.

When expressed as Cohen's d, the effect sizes for major depression and depression symptoms are similar: d = 0.37 and d = 0.27, respectively.

“Expressed in the Cohen’s d metric, the two effect sizes are similar: d = 0.37 for major depression and d = 0.27 for depression symptoms.”

May combine two claims.

In our meta-analysis, the odds ratio for ages 13-15 was 3.02 and declined to 2.69 for ages 16-19.

“The odds ratio for the 13–15 age group in our meta-analysis was already 3.02 and declined, not widened, to OR = 2.69 for ages 16–19.”

May combine two claims.

Categorical year-of-data-collection estimates for major depression were: 1991-1996 = 1.84 (k = 22); 1997-2002 = 1.88 (k = 134); 2003-2008 = 1.91 (k = 59); 2009-2014 = 2.17 (k = 85).

“To better understand this pattern, we created a categorical variable for year of data collection and obtained the following effect size estimates: 1991–1996= 1.84, k = 22; 1997–2002= 1.88, k = 134; 2003–2008= 1.91, k = 59; 2009–2014 = 2.17, k = 85.”

May combine two claims.

In the major depression meta-analysis, the youngest age studied was 12, and the odds ratio at that age was already 2.37.

“In the major depression meta-analysis, we could not examine the emergence of the gender difference given that the youngest age in the studies was 12, when the OR was already 2.37.”

May combine two claims.

Across different subgroups, every odds ratio for major depression diagnoses exceeded 1.0 and every effect size for depression symptoms was positive.

“That is, among different subgroups, all odds ratios for diagnoses were > 1.0, and all effect sizes for symptoms were positive.”

May combine two claims.

For depression symptoms, contraceptive prevalence, executive positions, and literacy ratio did not predict variation in effect size.

“Contraceptive prevalence, executive positions, and the literacy ratio did not predict variation in effect size.”

May combine two claims.

In follow-up analyses excluding outliers, diagnoses using ICD were significantly smaller than those using DSM-IV/DSM-IV-TR (QB = 6.86, p < .01) and DSM-III/DSM-III-R (QB = 9.60, p < .01).

“Follow up analyses, excluding outliers, indicated that diagnoses using the ICD were significantly smaller than both diagnoses using DSM-IV/DSM-IV-TR (QB = 6.86, p < .01) and DSM-III/ DSM- III-R TR (QB = 9.60, p < .01).”

May combine two claims.

In the symptom meta-analysis, the gender difference was trivial at ages 8-11, then increased steeply and peaked at age 16.

“In the symptom meta-analysis, the gender difference emerged in adolescence with a trivial gender difference for ages 8–11 (see Figure 3) and then a steep increase, reaching a peak in the gender difference at age 16.”

May combine two claims.

In follow-up tests, all pairwise comparisons were significantly different, with the smallest effect size for the BDI (d = 0.20) and the largest for scales other than the BDI and CES-D (d = 0.31).

“In follow-up tests, all pair-wise comparisons were significantly different from each other, such that the smallest effect size was for the BDI (d = 0.20) and the largest effect size was for scales other than the BDI and CES-D (d = 0.31).”

May combine two claims.

Income inequality had no relation to the gender effect size for diagnoses; it was negatively related for symptom measures but lost significance when outliers were excluded.

“However, there was no relation to the gender effect size for diagnoses; it was negatively related for symptom measures but lost significance when outliers were excluded.”

May combine two claims.

In both the diagnosis and symptom meta-analyses, differences among U.S. ethnic groups were not significant.

“In both the diagnosis and symptom meta-analyses, differences among ethnic groups were not significant.”

May combine two claims.

The symptom meta-analysis found no significant trend over time, while the diagnosis meta-analysis found a positive trend indicating growing gender differences.

“The symptom meta-analysis found no significant trend over time and the diagnosis meta-analysis found a positive trend, meaning that gender differences are growing larger.”

May combine two claims.

The finding of larger gender differences in nations with greater gender equity did not replicate in the symptom meta-analysis, so caution is needed in interpreting the diagnosis findings.

“These findings of larger gender differences in nations with greater gender equity did not replicate with the meta-analysis of depression symptoms, suggesting that caution is needed in interpreting the findings for diagnoses.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Other studies have found smaller gender differences in nations with more gender inequality for outcomes such as self-reports of personality traits and attitudes about mathematics.

“However, other studies have found smaller gender differences in nations with more gender inequality. This pattern has been found for outcomes such as self-reports of personality traits and attitudes about mathematics”

• cites Else-Quest et al., 2010 sourceNot checked — no full text of the source
• cites Wood & Eagly, 2012 sourceNot checked — no full text of the source

May combine two claims.

That meta-analysis was limited to the Children's Depression Inventory scale and samples of 8- to 16-year-old participants in the United States and Canada, with n = 43,916.

“it was limited to the Children’s Depression Inventory (CDI; Kovacs, 1985) scale and samples of 8- to 16-year-old participants in the United States and Canada (n = 43,916).”

• cites Twenge & Nolen-Hoeksema, 2002 sourceNot checked — no full text of the source
• cites Kovacs, 1985Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

Social-structural theory predicts that larger gender differences should be observed in nations with more gender inequality.

“The theory predicts that larger gender differences should be observed in nations with more gender inequality.”

• cites Eagly & Wood, 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Wood & Eagly, 2012 sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

Hankin and colleagues found that the gender difference in clinical depression emerged by ages 13 to 15 and then widened between ages 15 and 18.

“Hankin and colleagues (1998) found that the gender difference in clinical depression emerged by ages 13–15 and then widened between ages 15 and 18.”

• cites Hankin et al., 1998 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

According to social-structural theory, a society's division of labor by gender drives all other psychological gender differences.

“According to the theory, a society’s division of labor by gender drives all other psychological gender differences.”

• cites Eagly & Wood, 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Wood & Eagly, 2012 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Developmental psychopathology theories contend that multiple pathways to the gender difference in depression involve combinations and interactions of risk factors across multiple levels of analysis.

“Theories of developmental psychopathology contend that there are multiple pathways to the gender difference in depression involving combinations and interactions of risk factors that span multiple levels of analysis”

• cites Cicchetti & Rogosch, 2002 sourceNot checked — no full text of the source

May combine two claims.

In childhood, there is no gender difference or even a somewhat higher prevalence of depression among boys than girls.

“there is no gender difference or even a somewhat higher prevalence of depression among boys than girls in childhood”

• cites Avenevoli et al., 2008Not checked — no full text of the source
• cites Twenge & Nolen-Hoeksema, 2002 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

Evidence for social-structural theory exists for several psychological gender differences, including mate preferences, mathematics performance, and some aspects of sexuality.

“Evidence for this theory exists for several psychological gender differences, including mate preferences, mathematics performance, and some aspects of sexuality”

• cites Eagly & Wood, 1999Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Else-Quest et al., 2010 sourceNot checked — no full text of the source
• cites Petersen & Hyde, 2010 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

To the extent that women report higher rates of violent victimization, this may contribute to the gender difference in depression.

“To the extent that women report higher rates of violent victimization, this may contribute to the gender difference in depression”

• cites Koss et al., 1994 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

The female preponderance in depression is thought to emerge by ages 13 to 15.

“The female preponderance in depression is thought to emerge by ages 13–15”

• cites Hankin et al., 1998 sourceNot checked — no full text of the source
• cites Twenge & Nolen-Hoeksema, 2002 sourceNot checked — no full text of the source
• cites Wichstrøm, 1999 sourceNot checked — no full text of the source

May combine two claims.

Gender inequality is linked to discrimination against women, which may contribute to the gender difference in depression.

“gender inequality is linked to discrimination against women, which may contribute to the gender difference”

• cites Belle & Doucet, 2003 sourceNot checked — no full text of the source

Classified as probably not a substantial claim. May combine two claims.

Theories highlight that specific vulnerability factors come online at critical developmental periods in adolescence and/or interact with stressors in adolescence to produce the gender difference in adolescence.

“Theories highlight how specific vulnerability factors come on-line at critical developmental periods in adolescence and/or interact with stressors in adolescence to produce the gender difference in adolescence”

• cites Hyde et al., 2008bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Adolescents and adults with subthreshold depression symptoms and minor depression experience significant impairment and are at elevated risk for later major depression and suicidal behaviors.

“adolescents and adults with subthreshold depression symptoms and minor depression still experience significant impairment (e.g., Lewinsohn, Solomon, Seeley, & Zeiss, 2000) and are at elevated risk for later development of major depression and suicidal behaviors”

• cites Lewinsohn et al., 2000 sourceNot checked — no full text of the source
• cites Cuijpers et al., 2004Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

• cites Fergusson et al., 2005 sourceNot checked — no full text of the source

May combine two claims.

The extant literature indicates that the prevalence of major depression in the United States varies both by gender and by ethnicity.

“The extant literature indicates that the prevalence of major depression in the United States varies both by gender and by ethnicity”

• cites Breslau et al., 2006Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing).

One recent narrative review concluded that internalizing problems for girls increased from the late 20th century to the 21st century.

“One recent narrative review concluded that internalizing problems for girls increased from the late 20th century to the 21st century”

• cites Bor et al., 2014 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The authors followed MARS reporting standards.

“we follow MARS reporting standards”

• cites American Psychological Association Publications and Communications Board, 2008Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Classified as probably not a substantial claim.

To maintain independence of samples, when a sample was used in more than one article, the authors selected the article with the most complete data and/or the largest sample size.

“If a particular sample of participants was used in more than one article, which was often the case with these national datasets, to maintain independence of samples, we selected the article that had the most complete data (including information on moderator variables such as age and ethnicity) and/or the largest sample size.”

Classified as probably not a substantial claim.

For nationally representative longitudinal studies with multiple waves, only one wave of data was included to maintain independence of samples.

“For nationally representative longitudinal studies with multiple waves of data (e.g., Add Health), we included only one wave of data to maintain independence of samples.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Interrater agreement was κ = .87 for type of depression diagnosis and κ = 1.00 for time span of the depression diagnosis.

“type of depression diagnosis (depressive episode versus depressive disorder), κ = .87; and (4) the time span for the depression diagnosis (current, 1 month, 6 months, 12 months, lifetime), κ = 1.00.”

May combine two claims.

All datasets were nationally representative, which minimizes sampling bias.

“First, all datasets were nationally representative, indicating that any bias in sampling is minimized.”

May combine two claims.

For the diagnosis meta-analysis, author-provided data were obtained for 24 of 65 articles (37%), and 167 of 300 effect sizes (56%) were based on obtained rather than published data.

“For the meta-analysis on depression diagnoses, we received data from authors for 24 (37%) of the 65 articles, such that 167 (56%) of the 300 effect sizes were based on obtained rather than published data.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The intimate-partner violence and sexism ideals measures were not considered further because they were available for only a minority of effect sizes.

“Although they were theoretically interesting, the intimate-partner violence and sexism ideals measures proved unsatisfactory because they were available for only a minority of the effect sizes in analyses. Those two measures are therefore not considered further.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Data were entered into the effect size calculator by two individuals to ensure accuracy.

“Data were entered in the calculator by two individuals to ensure accuracy.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Interrater agreement for article focus was κ = .84.

“we coded the focus of the article (gender, depression, other), κ = .84.”

Classified as probably not a substantial claim.

Hedges and Nowell argued that the strongest scientific evidence about gender differences comes from larger studies based on representative samples of populations, not small convenience samples.

“They argued that the strongest scientific evidence about gender differences does not come from small studies of convenience samples, but instead comes from larger studies based on representative samples of populations.”

• cites Hedges & Nowell, 1995 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Abstracts were included only if they explicitly mentioned "nationally representative."

“We included abstracts that explicitly mentioned “nationally representative.””

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Interrater agreement was κ = 1.0 for the type of measure used to assess depression and for the diagnostic manual used.

“the type of measure used to assess depression diagnoses (diagnostic interview; e.g., CIDI, DIS) or depression symptoms (self-report measure; e.g., CES-D, BDI), κ = 1.0; (2) the manual used to diagnose depression (DSM-IV or DSM-IV-TR5; DSM-III or DSM-III-R; or ICD-10), κ = 1.0”

May combine two claims.

For the symptom meta-analysis, author-provided data were obtained for 71 of 95 articles (75%), and 357 of 413 effect sizes (86%) were based on obtained data.

“For the meta-analysis on depression symptoms, we received data from authors for 71 (75%) of the 95 articles, such that 357 (86%) of the 413 effect sizes were based on obtained data.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Interrater agreement was high for age (r = .99), country (κ = 1.00), U.S. ethnic group (κ = 1.00), and year of data collection (r = 1.00).

“age of the participants (measured on a continuous scale, using the reported mean or the midpoint of the age range), r = .99”

• Also stated in: § Coding of Studies · plain

  “country, κ (kappa) = 1.00”

• Also stated in: § Coding of Studies · plain

  “predominant (≥ 85%) ethnic group of participants (Caucasian, African American, Asian American, Hispanic, Native American, other, mixed, or unreported), κ = 1.00; and (d) year of data collection, r = 1.00.”

Classified as probably not a substantial claim. May combine two claims.

A meta-analysis of gender differences in self-esteem found a small difference favoring Caucasian males over Caucasian females, d = 0.20, but no gender difference for African American samples, d = −0.04.

“a meta-analysis of gender differences in self-esteem found a small difference favoring Caucasian males over Caucasian females, d = 0.20, but no gender difference for African American samples, d = −0.04”

• cites Kling et al., 1999 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing).

The authors contacted study authors for whom email addresses could be found to obtain missing data.

“all authors of the study for whom we could find email addresses from the article, the Web directory of the authors’ academic institution, or a Google search, were contacted.”

Classified as probably not a substantial claim.

Studies were excluded if their depression symptom measure did not meet criteria: at least 3 items, self-report, Cronbach's alpha ≥ .70 if provided, and a valid and reliable measure based on previously published research.

“Studies were excluded if their measurement of depression symptoms did not meet the following criteria: 1) minimum of 3 items; 2) self-report; 3) Cronbach’s alpha ≥.70 (if provided); and 4) valid and reliable measure of depression based on previously published research”

Classified as probably not a substantial claim. May combine two claims.

Studies were excluded if they were not based on national probability sampling; only population-based surveys representative of the country were included.

“we excluded studies that were not based on national probability sampling. In other words, we included only population-based surveys representative of the country.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Thirty studies were double coded to compute interrater agreement.

“We double coded 30 studies to compute interrater agreement.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Data for the economic and gender equity indicators were not available for all nations, so sample sizes varied across regression analyses depending on the indicator.

“For each of the economic and gender equity indicators, data are not available for all nations. Therefore, sample sizes vary in regression analyses, depending on the indicator.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Raw effect sizes were corrected for small-sample bias, and most correction factors were close to 1.0 because sample sizes were large.

“Raw effect sizes were corrected for bias (i.e., the upward-bias of effect sizes among small samples; Hedges, 1981); however, most correction factors were close to 1.0 given the large sample sizes.”

• cites Hedges, 1981 sourceNot checked — no full text of the source

May combine two claims.

Because most studies were published on depression rather than on gender differences, publication bias favoring gender differences should not be a problem.

“This indicates that most studies were published on the basis of the work on depression and not on the basis of a gender difference, so publication bias in the direction of gender differences should not be a problem.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The Hedges and Nowell strategy has been used in other meta-analyses on gender differences in cognitive abilities.

“The Hedges and Nowell strategy has since been used in other meta-analyses on gender differences in cognitive abilities”

• cites Else-Quest et al., 2010 sourceNot checked — no full text of the source
• cites Reilly et al., 2015 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Studies were excluded if their measurement of depression diagnoses did not include a diagnostic interview with the participant.

“Studies were excluded in the processing of articles if their measurement of depression diagnoses did not include a diagnostic interview with the participant.”

Probably not a substantial claim: may not pass S3 (load-bearing).

All analyses were run with the full dataset and then re-run with outliers excluded.

“All analyses were run with the full data set and then re-run excluding outliers.”

Classified as probably not a substantial claim. May combine two claims.

Stressors tied to major life transitions are spread more evenly across adulthood, leading to more even depression rates across age for both genders and a stable gender gap.

“The result is that stressors attached to these transitions are spread out more evenly across adulthood, leading to more even rates of depression across age for both women and men, and a stable gender gap.”

May combine two claims.

Although the range of ORs from 1.84 to 2.17 is not large, the positive relationship between year of data collection and the gender difference in major depression is clear.

“Thus, although the range of OR = 1.84 to 2.17 is not great, the positive relationship is clear.”

May combine two claims.

Our results provide powerful evidence that the gender difference in depression emerges earlier than previously thought, by at least age 12 for diagnoses and at age 12 for symptoms.

“Taken together, our results provide powerful evidence that the gender difference in depression emerges earlier than previously thought (by at least age 12 for diagnoses, at age 12 for symptoms), which has important implications for the timing of preventive interventions.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Hankin and colleagues (1998) found that the gender gap in major depression emerged between ages 13 and 15 and widened between ages 15 and 18.

“These results differ from those of Hankin and colleagues (1998), who found that that the gender gap in major depression emerged between ages 13 and 15 and then widened between ages 15 and 18.”

• cites Hankin and colleagues (1998)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

The adult pattern is a new finding and should be robust because it comes from large-scale meta-analyses and was consistent across diagnosis and symptom measures.

“This pattern is a new finding and should be robust because it is based on large-scale meta-analyses and was consistent across both diagnosis and symptom measures.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Individuals high in negative affectivity and low in positive affectivity are vulnerable to later depression.

“In particular, individuals who are high in negative affectivity and low on positive affectivity are vulnerable to later depression.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

According to the cognitive vulnerability-stress model, the adolescent gender difference in depression may be explained by higher negative cognitive style in girls, higher stress in girls, or both.

“According to this model, the gender difference in depression in adolescence may be accounted for by (a) higher levels of negative cognitive style in girls than boys beginning in early adolescence; (b) higher levels of stress for girls than boys beginning in early adolescence; or (c) both.”

May combine two claims.

Females consistently show higher levels of major depression and depression symptoms than males.

“This emphasizes the consistency with which females have higher levels major depression and depression symptoms than males.”

Probably not a substantial claim: may not pass S3 (load-bearing).

One theoretical account holds that early puberty is disadvantageous for girls but not boys for outcomes such as depression.

“According to one theoretical account, early puberty is disadvantageous for girls but not boys, for outcomes such as depression (Ge, Conger, & Elder, 2001).”

• cites Ge, Conger, & Elder, 2001 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The difference between our findings and earlier meta-analytic findings may be due to more recent studies, more nations, or multiple symptom measures.

“The difference in findings may be due to the greater recency of many of our studies, the greater number of nations, or the inclusion of multiple measures of depression symptoms.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Depression is a recurrent disorder.

“A third possibility results from the observation that depression is a recurrent disorder (e.g., Lewinsohn, Rohde, Seeley, Klein, & Gotlib, 2000).”

• cites Lewinsohn, Rohde, Seeley, Klein, & Gotlib, 2000 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Men may be less likely than women to develop depression, but depressed men are still distressed and impaired.

“Men may be less likely to develop depression than women; however, this does not mean that depressed men are not distressed and impaired.”

May combine two claims.

Gender equity measures were missing for some nations, so the analyses did not capture patterns among all included nations.

“Second, gender equity measures were missing for some nations; thus, the analyses did not capture patterns among all nations included.”

Probably not a substantial claim: may not pass S3 (load-bearing).

One theoretical account holds that temperament, present from infancy and early childhood, predicts later depression.

“According to one theoretical account, temperament, present from infancy and early childhood, predicts later depression (summarized by Hyde et al., 2008b).”

• cites Hyde et al., 2008bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

In the 21st century, women make up 47% of the U.S. labor force, compared with 30% in 1950.

“For example, in the 21st century women constitute 47% of the U.S. labor force, compared to 30% in 1950 (Costello et al., 2003).”

• cites Costello et al., 2003Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Negative cognitive style may not become a stable trait until ages 9.5 to 12.5, and the cognitive vulnerability-stress interaction may not reliably predict depression until ages 13.5 to 14.5.

“Research suggests that negative cognitive style may not emerge as a stable trait until ages 9.5 to 12.5, and the cognitive vulnerability-stress interaction does not become a reliable predictor of depression until ages 13.5 to 14.5, i.e., in early adolescence (Cole et al., 2008).”

• cites Cole et al., 2008 sourceNot checked — no full text of the source

May combine two claims.

The gender difference in depression is created at least in part by early-puberty girls, through processes such as encountering more peer sexual harassment than boys and on-time girls.

“Thus the gender difference in depression is created at least in part by girls who go through puberty early, because of any of several processes, such as early-puberty girls encountering more peer sexual harassment than boys and on-time girls (Lindberg, Grabe, & Hyde, 2007).”

• cites Lindberg, Grabe, & Hyde, 2007 sourceNot checked — no full text of the source

May combine two claims.

Symptom assessments typically cover a short period such as a week, while diagnoses aggregate over longer periods such as a year.

“For example, the typical assessment of symptoms often represents a short period of time, such as a week, whereas diagnoses involve aggregation over longer periods, often a year (Haeffel et al., 2003).”

• cites Haeffel et al., 2003 sourceNot checked — no full text of the source

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Most first cases of depression may occur in adolescence and then recur at variable times in adulthood because of the uneven timing of major stressors.

“As such, it may be that most first cases of depression occur in adolescence, and then recur at variable times in adulthood, as a result of the uneven timing of major stressors.”

May combine two claims.

Theories guided by developmental psychopathology have focused on the emergence of the gender difference in adolescence but have not attended to development across adulthood.

“As noted earlier, theories guided by developmental psychopathology have focused on explaining the emergence of the gender difference in adolescence (summarized by Hyde et al., 2008b), but did not attend to development across adulthood.”

• cites Hyde et al., 2008bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

An odds ratio of 1.95 is a medium rather than large effect size, but it still represents a health disparity.

“An OR of 1.95 is a medium, not a large, effect size, yet it is still a health disparity.”

May combine two claims.

Another theoretical account rests on cognitive vulnerability-stress models of depression, which have been well supported in college student and adult samples.

“Another theoretical account rests on cognitive vulnerability-stress models of depression, which have been well supported in samples of college students and adults (summarized by Hyde et al., 2008b).”

• cites Hyde et al., 2008bNot matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Prevalence would remain stable across adulthood for both women and men, producing a stable gender gap created by the original adolescent gap.

“Again, prevalence would remain stable across adulthood for both women and men, leading to a stable gender gap, created by the original gender gap in adolescence.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Depression is a global health priority.

“Given that depression is a global health priority (World Health Organization, 2016), it is imperative to understand disparities in depression and which subgroups are most in need of services.”

• cites World Health Organization, 2016Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

Assuming equal numbers of males and females, an odds ratio of 2.0 can result from a 9.05% female and 4.85% male prevalence, or from an 18% female and 10.1% male prevalence.

“For example, assuming equal numbers of males and females, an OR = 2.0 can result from a 9.05% prevalence in females and a 4.85% prevalence in males, or from an 18% prevalence in females and a 10.1% prevalence in males.”

May combine two claims.

Evidence indicates high reliability between clinicians and lay interviewers; for major depressive disorder, kappa was 0.97 with 99.7% agreement.

“Results indicated high agreement between the two; for major depressive disorder, kappa = 0.97, with 99.7% agreement between the two sources (see also Brugha, Nienhuis, Bagchi, Smith, & Meltzer, 1999).”

• cites Wittchen, Robins, Cottler, Sartorius, Burke, & Regier, 1991 sourceNot checked — no full text of the source
• cites Brugha, Nienhuis, Bagchi, Smith, & Meltzer, 1999 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

The authors made extensive efforts to obtain data for as many national data sets as possible by conducting additional computerized searches and contacting authors.

“We also made extensive efforts to obtain data for as many national data sets as possible by conducting additional computerized searches and contacting authors.”

May combine two claims.

Mounting a nationally representative study is a major, costly undertaking that should almost certainly result in multiple publications, at least one in English, so such studies should have been detected.

“Moreover, mounting a study based on a nationally representative sample is a major, costly undertaking that should almost certainly result in multiple publications, at least one of them in English, so we should have detected such studies.”

May combine two claims.

Gender differences were larger in nations with more gender equity, a finding consistent across three indicators.

“That is, perhaps counterintuitively, gender differences were larger in nations with more gender equity, a finding that was consistent across three indicators.”

• AssumptionRests on an assumption: The nation-level indicators of gender equity and national wealth validly measure those constructs.

Probably not a substantial claim: may not pass S3 (load-bearing).

Because the study focused on large nationally representative datasets, diagnostic interviews were conducted by trained lay interviewers rather than skilled clinicians, as is standard in such epidemiological studies.

“Second, focusing on large, nationally representative datasets meant that the diagnostic interviews for major depression were conducted not by skilled clinicians, but instead by trained lay interviewers, as is standard practice in these large-scale, epidemiological studies.”

Classified as probably not a substantial claim. May combine two claims.

Both diagnostic measures and symptom questionnaire measures rely on self-reports from participants.

“Third, both diagnostic measures and symptom questionnaire measures rely on self-reports from participants.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Some primary studies have found patterns similar to ours regarding gender differences and gender equity for outcomes such as self-construals and emotion.

“Yet, as noted in the introduction, some primary studies have found similar patterns to ours regarding gender differences and gender equity, for outcomes such as self-construals (Guimond et al., 2007) and emotion (Fischer, Rodriguez Mosquera, van Vianen, & Manstead, 2004).”

• cites Guimond et al., 2007 sourceNot checked — no full text of the source
• cites Fischer, Rodriguez Mosquera, van Vianen, & Manstead, 2004 sourceNot checked — no full text of the source

Classified as probably not a substantial claim.

Guimond and colleagues (2007) proposed that gender differences in self-construals are larger when individuals make intergroup social comparisons.

“Guimond and colleagues (2007) proposed that gender differences in self-construals are larger in contexts in which individuals make intergroup social comparisons, e.g., when they compare themselves to an out-group such as the other gender.”

• cites Guimond and colleagues (2007)Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

These results suggest that women worldwide are at significantly greater risk of depression diagnoses and symptoms than men, and that adolescent girls are at greatest risk.

“These results suggest that women are at significantly greater risk of depression diagnoses and symptoms compared to men worldwide, and that adolescent girls are at the greatest risk.”

May combine two claims.

By synthesizing nationally representative studies with over 1.7 million participants each, the two meta-analyses provide a comprehensive quantitative review of gender differences in major depression diagnoses and depression symptoms across the lifespan.

“By synthesizing nationally representative studies with data from over 1.7 million participants spanning the globe in each of the two meta-analyses, we have provided a comprehensive quantitative review of data on gender differences in major depression diagnoses and depression symptoms across the lifespan.”

• AssumptionRests on an assumption: The nationally representative samples included in the meta-analyses accurately represent their target national populations.

May combine two claims.

In higher gender-equity nations, where males and females interact more, intergroup comparisons are likelier, leading to larger gender differences.

“It is precisely in higher gender-equity nations, where males and females interact more (e.g., girls are more literate because they have more equal access to schooling with boys), that intergroup comparisons are likelier, leading to larger gender differences.”

Classified as probably not a substantial claim.

The database search included only articles in English.

“Fourth, our database search included only articles in English.”

May combine two claims.

The ethnic-group analyses were based on small numbers of studies, so power to detect ethnic-group differences was limited.

“These analyses, however, were based on small numbers of studies. Therefore, power to detect ethnic-group differences was limited.”

May combine two claims.

In low-gender-equity nations, intergroup interactions are often restricted, and gender differences on variables such as depression are smaller because comparisons are mostly intragroup.

“In low-gender-equity nations, intergroup interactions are often greatly restricted, and gender differences on a variable such as depression are smaller due to mostly intragroup comparisons.”

May combine two claims.

The authors did not have sufficient data to examine gender differences in major depression for children younger than age 12, and had only 2 samples for age 12, even though the minimum age criterion was 7.

“First, we did not have sufficient data to examine gender differences in major depression for children younger than age 12 (and we had only 2 samples for age 12), even though the minimum age criterion was 7.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

If there are gender differences in willingness to disclose symptoms, then the resulting data may be biased.

“If there are gender differences in willingness to disclose symptoms, then the resulting data may be biased.”

May combine two claims.

The difference between the high-income and other-nation odds ratios is small, with overlapping 95% confidence intervals, suggesting it is not a major effect.

“However, the difference between these odds ratios is small, and the 95% confidence intervals overlap, suggesting that this difference is not a major effect.”

Probably not a substantial claim: may not pass S1 (asserted), S3 (load-bearing).

One early review concluded that the gender difference in depression is a real difference and not a measurement artifact.

“One early review concluded that the gender difference in depression is a real difference and not a measurement artifact (Weissman & Klerman, 1977).”

• cites Weissman & Klerman, 1977Not matched

  Couldn’t be matched to a paper in Crossref, so it can’t be checked.

May combine two claims.

The authors believe that the English-only search did not lead to the omission of nations in which English is not the predominant language, for two reasons.

“We believe that this did not lead to the omission of nations in which English is not the predominant language for two reasons.”

May combine two claims.

Depression is a recurring disorder, so an episode in adolescence can predispose an individual to later episodes.

“The emphasis on adolescents is particularly important because depression is a recurring disorder, so an episode in adolescence can predispose the individual to later episodes (e.g., Lewinsohn, Rohde, Seeley, Klein, & Gotlib, 2000).”

• cites Lewinsohn, Rohde, Seeley, Klein, & Gotlib, 2000 sourceNot checked — no full text of the source

May combine two claims.

The gender difference in depression is a major health disparity, especially in adolescence, but the magnitude of the difference indicates that depression in males should not be overlooked.

“The gender difference in depression represents a major health disparity, especially in adolescence, yet the magnitude of the difference indicates that depression in males should not be overlooked.”

Not yet classified.

Overall, 76% of symptom effect sizes and 56% of diagnosis effect sizes were based on data supplied by researchers rather than published articles, reducing concerns about publication bias.

“Overall, 76% of the effect sizes for the symptom meta-analysis and 56% of the effect sizes for the diagnosis meta-analysis were based on data supplied by researchers, and not published in articles, reducing concerns about publication bias.”