Results everyone can understand: A review of common language effect size indicators to bridge the research-practice gap.

Mastrich, Zachary H.; Hernandez, Ivan

doi:10.1037/hea0001112

Cited by 12 publications

(7 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For raw data, all values that deviated more than 3.5 SD from the mean were classified as outliers and removed 6 . To put findings into perspective, we additionally report a range of natural language interpretations that are more easily interpreted than standardized mean effect sizes (Mastrich & Hernandez, 2021). First, Cohen’s U 3, a measure of nonoverlap, indicates which percentage of Population A is surpassed by the upper half of Population B (Cohen, 1988).…”

Section: Methodsmentioning

confidence: 99%

Sex drive: Theoretical conceptualization and meta-analytic review of gender differences.

Frankenbach¹,

Weber²,

Loschelder³

et al. 2022

Psychological Bulletin

View full text Add to dashboard Cite

Few spheres in life are as universally relevant for (almost) all individuals past puberty as sexuality. One important aspect of sexuality concerns individuals’ sex drive—their dispositional sexual motivation. A vigorous scientific (and popular) debate revolves around the question of whether or not there is a gender difference in sex drive. Several theories predict a higher sex drive in men compared to women, with some theories attributing this difference to biased responding rather than true differences. Currently, there is little consensus on how to conceptualize sex drive, nor does a quantitative summary of the literature exist. In this article, we present a theory-driven conceptualization of sex drive as the density distribution of state sex drive, where state sex drive is defined as momentary sexual motivation that manifests in sexual cognition, affect, and behavior. We conduct a comprehensive meta-analysis of gender differences in sex drive based on 211 studies, 856 effect sizes, and 621,463 persons. The meta-analysis revealed a stronger sex drive in men compared to women, with a medium-to-large effect size, g = 0.69, 95% CI [0.58, 0.81]. Men more often think and fantasize about sex, more often experience sexual affect like desire, and more often engage in masturbation than women. Adjustment for biased responding reduced the gender difference (g = 0.54). Moderation analyses suggest that the effect is robust and largely invariant to contextual factors. There was no evidence of publication bias. The discussion focuses on validity considerations, limitations, and implications for psychological theory and people’s everyday lives.

show abstract

Section: Methodsmentioning

confidence: 99%

Sex drive: Theoretical conceptualization and meta-analytic review of gender differences.

Frankenbach¹,

Weber²,

Loschelder³

et al. 2022

Psychological Bulletin

View full text Add to dashboard Cite

show abstract

“…SD AIIP = 36.60), d = 0.18. The effect size suggests a 5% difference in the probability that a randomly selected AI-IP item will be less readable than a randomly selected IPIP item, compared to another randomly selected IPIP item (Mastrich & Hernandez, 2021). From an absolute perspective, the Flesch Reading Ease scoring guides indicate that a readability score of 68.72 is on the margin between "standard" (60-69) and "fairly easy" (70-79), implying that the AI-IP items are, on average, not difficult to read.…”

Section: Resultsmentioning

confidence: 99%

“…The AI‐IP on average had items that were approximately 1 word shorter ( M words = 5.19; SD = 1.76) than the IPIP ( M words = 6.74; SD = 3.38), d = 0.58. In common language effect size terms, there is a 65.79% chance that a randomly selected AI‐IP item will be shorter than a randomly selected item from the IPIP (Mastrich & Hernandez, 2021). Therefore, although there is a practical difference in item length, it may be desirable to have more concise items, if they demonstrate equivalent validity.…”

Section: Resultsmentioning

confidence: 99%

“…While, on average, the items in the original IPIP are more readable ( M readability = 74.88) than the AI‐IP ( M readability = 68.72), this difference is also minimal compared to the vast variation in readability within each scale ( SD IPIP = 30.24; SD AIIP = 36.60), d = 0.18. The effect size suggests a 5% difference in the probability that a randomly selected AI‐IP item will be less readable than a randomly selected IPIP item, compared to another randomly selected IPIP item (Mastrich & Hernandez, 2021). From an absolute perspective, the Flesch Reading Ease scoring guides indicate that a readability score of 68.72 is on the margin between “standard” (60–69) and “fairly easy” (70–79), implying that the AI‐IP items are, on average, not difficult to read.…”

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

The AI‐IP: Minimizing the guesswork of personality scale item development through artificial intelligence

Hernandez

Nie

2022

Personnel Psychology

Self Cite

View full text Add to dashboard Cite

We propose a framework for integrating various modern natural language processing (NLP) models to assist researchers with developing valid psychological scales.Transformer-based deep neural networks offer state-ofthe-art performance on various natural language tasks. This project adapts the transformer model GPT-2 to learn the structure of personality items, and generate the largest openly available pool of personality items, consisting of one million new items. We then use that artificial intelligencebased item pool (AI-IP) to provide a subset of potential scale items for measuring a desired construct. To better recommend construct-related items, we train a paired neural network-based classification BERT model to predict the observed correlation between personality items using only their text. We also demonstrate how zero-shot models can help balance desired content domains within the scale. In combination with the AI-IP, these models narrow the large item pool to items most correlated with a set of initial items. We demonstrate the ability of this multimodel framework to develop longer cohesive scales from a small set of construct-relevant items. We found reliability, validity, and fit equivalent for AI-assisted scales compared toThis is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

show abstract

“…In order to ascertain the effect size of the JD on T2D, we computed the nonparametric common language effect size (CLES) for significant variables. The value represents the probability that a value chosen randomly from JD7 will differ from a value chosen randomly from BL [ 28 , 29 ]. We performed a post hoc sample size calculation using the primary endpoint of SDNN and using a 30% change this endpoint, with the probability of a type 1 error ( α = 0.05) and type 2 error ( β = 0.05); the required sample size of n = 7 would be needed to yield a power of 0.95.…”

Section: Methodsmentioning

confidence: 99%

A Nonrandomized Trial of the Effects of Passive Simulated Jogging on Short-Term Heart Rate Variability in Type 2 Diabetic Subjects

Adams

López

Banderas³

et al. 2023

Journal of Diabetes Research

View full text Add to dashboard Cite

Background. Diabetes mellitus has reached global epidemic proportions, with type 2 diabetes (T2DM) comprising more than 90% of all subjects with diabetes. Cardiovascular autonomic neuropathy (CAN) frequently occurs in T2DM. Heart rate variability (HRV) reflects a neural balance between the sympathetic and parasympathetic autonomic nervous systems (ANS) and a marker of CAN. Reduced HRV has been shown in T2DM and improved by physical activity and exercise. External addition of pulses to the circulation, as accomplished by a passive simulated jogging device (JD), restores HRV in nondiseased sedentary subjects after a single session. We hypothesized that application of JD for a longer period (7 days) might improve HRV in T2DM participants. Methods. We performed a nonrandomized study on ten T2DM subjects (age range 44-73 yrs) who were recruited and asked to use a physical activity intervention, a passive simulated jogging device (JD) for 7 days. JD moves the feet in a repetitive and alternating manner; the upward movement of the pedal is followed by a downward movement of the forefoot tapping against a semirigid bumper to simulate the tapping of feet against the ground during jogging. Heart rate variability (HRV) analysis was performed using an electrocardiogram in each subject in seated posture on day 1 (baseline, BL), after seven days of JD (JD7), and seven days after discontinuation of JD (Post-JD). Time domain variables were computed, viz., standard deviation of all normal RR intervals (SDNN), standard deviation of the delta of all RR intervals (SDΔNN), and the square root of the mean of the sum of the squares of differences between adjacent RR intervals (RMSSD). Frequency domain measures were determined using a standard Fast Fourier spectral analysis, as well as the parameters of the Poincaré plots (SD1 and SD2). Results. Seven days of JD significantly increased SDNN, SDΔNN, RMSSD, and both SD1 and SD2 from baseline values. The latter parameters remained increased Post-JD. JD did not modify the frequency domain measures of HRV. Conclusion. A passive simulated jogging device increased the time domain and Poincaré variables of HRV in T2DM. This intervention provided effortless physical activity as a novel method to harness the beneficial effects of passive physical activity for improving HRV in T2DM subjects.

show abstract

Results everyone can understand: A review of common language effect size indicators to bridge the research-practice gap.

Cited by 12 publications

References 49 publications

Sex drive: Theoretical conceptualization and meta-analytic review of gender differences.

Sex drive: Theoretical conceptualization and meta-analytic review of gender differences.

The AI‐IP: Minimizing the guesswork of personality scale item development through artificial intelligence

A Nonrandomized Trial of the Effects of Passive Simulated Jogging on Short-Term Heart Rate Variability in Type 2 Diabetic Subjects

Contact Info

Product

Resources

About