/Factor_Analysis

Factor analytics techniques employed in R, including EFA and CFA, to analyze Martin & Doris's (2003) research on the development of a psychometric instrument measuring individual styles of humor.

Primary LanguageR

Factor_Analysis

The primary project in this repository, humor_factoranalysis.R, analyzes questionnaire data collected from Martin & Doris's (2003) research on the development of a humor measurement scale called the Humor Styles Questionnaire (HSQ). The survey items that were generated to comprise the HSQ were born out of an extensive review on literature about how and why we use humor when interacting with others. There were 32 survey items in total that the reseachers chose to represent what they believed to be four underlying dimensions that describe individual differences in one's style of humor. Those four dimensions were:

  1. Affiliative humor: use of humor to facilitate relationships with others and to put others at ease
  2. Self-enhancing humor: one's ability to use humor to protect oneself against stress and adversity
  3. Aggressive humor: use of humor to manipulate others through mockery and shaming; making fun of someone else
  4. Self-defeating humor: amusing others by making onself the object of jokes

For this repository, I accessed survey data results from the original study that were made avaialable via the Open-Source Psychometrics Project. Without being familiar with previous research and theory on humor, I employ Exploratory Factor Analysis (EFA) to determine the underlying multiple dimensions of humor captured by the HSQ survey instrument. After assessing EFA model fit statistics, I conduct Confirmatory Factor Analysis (CFA) and assess CFA model fit statistics to determine whether certain items from the Humor Styles Questionnaire should be removed. Later I compare relative model fit statistics between the EFA and CFA models to determine the best factor structure for the data.

(n = 1,071)

The other projects in this repository provide code for data cleaning and factor analysis of other psychometric scales.

alt text

alt text

alt text