Understanding logistic regression analysis

Sperandei, Sandro

doi:10.11613/bm.2014.003

Cited by 872 publications

(588 citation statements)

References 10 publications

Supporting

Mentioning

569

Contrasting

Unclassified

Order By: Relevance

“…Consequently, the specific model that was used was a binary logistic model due to the fact that the dependent variable was binomial (Khan & Raeside, 1997;Sperandei, 2014;Khan & Rahman, 2016). The outcome of interest, matric completion, was constructed using the information on the highest level of education attained.…”

Section: Resultsmentioning

confidence: 99%

Effect of Early Sexual Debut on High School Completion in South Africa

2017

View full text Add to dashboard Cite

Early sexual debut is of major concern because it is a correlate for health and economic shocks experienced in adulthood. In South Africa, this concern has provided impetus for research directed at the HIV and AIDS epidemic, teenage pregnancy and the effect of adolescent sexual behaviour on persistence in school. Of interest to the present study is high school completion, which is a well-established empirical barometer of adult socioeconomic opportunities. Using data from the five waves of the Cape Area Panel Study (CAPS), this paper examines the association between sexual behaviours initiated in pre- and early adolescence and high school completion rates. The CAPS study is a longitudinal survey that was designed to investigate young people's (aged 14-22 years) educational attainment and sexual behaviours in Cape Town, South Africa. The sample was constituted from 3213 individuals who had initiated sex during their teenage years and the analysis was undertaken when the youngest cohort was aged 21, an age at which they should have completed high school if they were on time. Logistic regression models were fitted separately for males and females. Overall, the results reveal that early sexual debut is correlated with long-term negative educational outcomes. Individuals who experience early sexual debut are less likely to complete high school than their counterparts who make their sexual debut later on in life. This effect is worse for Africans, who also disproportionately have an earlier sexual debut than other race groups. Apart from race however, the findings also reinforce the effect of other demographic factors on high school completion, namely, place of residence and family socioeconomic status as measured by parental education and household income. Hence, early sexual debut adds another layer of inequality and worsens the plight of Africans, females, those living in rural areas and those who come from low-income families.

show abstract

Section: Resultsmentioning

confidence: 99%

Effect of Early Sexual Debut on High School Completion in South Africa

2017

View full text Add to dashboard Cite

show abstract

“…As is common when building a logistic regression model, we used a less stringent threshold for significance when deciding which variables to enter into the multivariate model, since the true size and significance of the effect of certain variables may be masked by potential confounding variables, and revealed after adjusting for those variables in the model. 23 No inferences are drawn from these univariate analyses; they are simply a screening to determine which independent variables are included in the model. Any variable with an adjusted odds ratio with a 95% confidence interval that did not cross 1 in the multivariate model was considered statistically significant.…”

Section: Methodsmentioning

confidence: 99%

Initial symptom burden predicts duration of symptoms after concussion

Meehan

O’Brien

Geminiani

et al. 2016

Journal of Science and Medicine in Sport

View full text Add to dashboard Cite

Objectives To determine which variables predict prolonged (>28 days) duration of symptoms after a concussion. Design We conducted a prospective cohort study of adult (>18yo) patients cared for in a specialty concussion clinic. Methods Symptoms were assessed using the Post-Concussion Symptom Scale (PCSS) developed at the 3rd International Conference on Concussion in Sports. Possible predictors including age, sex, loss of consciousness, amnesia, history of prior concussion, prior treatment for headaches, history of migraines, and family history of concussions, were measured by self-report. We recorded a PCSS score at each clinical visit and defined time to symptom resolution as the number of days between the date of injury and date of last symptoms. Results Of 64 adult patients included in the study, 53.3% were male; 20.3% reported experiencing a loss of consciousness at the time of injury while 23.4% reported amnesia. Patients ranged in age from 18 to 27 years (mean 21 ± 2 years). Most concussions (92.2%) occurred during sports. The mean initial PCSS score for those suffering symptoms for longer than 28 days was significantly higher than those who symptoms resolved within 28 days (42.5 vs. 19.2, p < 0.01). Of all potential predictor variables, only the initial PCSS score was independently associated with the odds of symptoms lasting longer than 28 days (aOR 1.037; 95% CI 1.011, 1.063). Conclusions Among adult patients with concussions, those with a higher symptom burden after injury have an increased odds of suffering from prolonged symptoms. Other potential predictor variables are not associated with the risk of prolonged recovery.

show abstract

“…As the significance of variables may change when considered in relation to other variables, any potential predictor variable with a P value of less than or equal to .25 was included in the initial regression model. 41 Potential predictors were baseline and short-term outcome variables and duration (in months) to long-term follow-up. Variables that were not considered significant were removed one by one, starting with the least significant variable, until a final model was achieved.…”

Section: Logistic Regressionmentioning

confidence: 99%

Long-Term Clinical Outcomes and Factors That Predict Poor Prognosis in Athletes After a Diagnosis of Acute Spondylolysis: A Retrospective Review With Telephone Follow-up

Selhorst¹,

Fischer²,

Graft³

et al. 2016

J Orthop Sports Phys Ther

View full text Add to dashboard Cite

® T t STUDY DESIGN:Retrospective review with telephone follow-up. T t BACKGROUND:Acute spondylolytic injuries have a dramatic impact on the young athlete. Excellent short-term clinical outcomes have been observed, but not enough is known about longterm clinical outcomes. T t OBJECTIVES:(1) To report long-term clinical outcomes for patients diagnosed with acute spondylolysis, and (2) to assess the prognostic ability of retrospective variables for long-term outcomes. T t METHODS: Patients from 2010 through 2013were retrospectively reviewed to identify patients and to obtain demographic, baseline, and shortterm outcomes. Long-term follow-up data were collected by telephone from patients diagnosed with acute spondylolysis to assess recurrence rate of low back pain, perceived outcome, pain, and functional ability. Patients were categorized as having a good or poor long-term outcome based on these measures. Logistic regression analysis was performed to assess the prognostic ability of the retrospective variables for long-term outcomes 3.4 years (range, 1.5-5.6 years) after treatment. T t RESULTS:One hundred twenty-one (71.6%) patients completed the follow-up questionnaire (48 female; mean age at baseline, 14.4 years). At follow-up, 81 (66.9%) patients were able to maintain their same or a higher level of sport. Recurrence of significant symptoms was reported by 55 (45.5%) patients, with 41 (33.9%) requiring medical treatment. The final logistic regression model revealed that female sex, adverse reaction during care, and multilevel injury were significant predictors of poor long-term outcome (R 2 = 0.22).

show abstract

Understanding logistic regression analysis

Cited by 872 publications

References 10 publications

Effect of Early Sexual Debut on High School Completion in South Africa

Effect of Early Sexual Debut on High School Completion in South Africa

Initial symptom burden predicts duration of symptoms after concussion

Long-Term Clinical Outcomes and Factors That Predict Poor Prognosis in Athletes After a Diagnosis of Acute Spondylolysis: A Retrospective Review With Telephone Follow-up

Contact Info

Product

Resources

About