week 9: qualitative data analyses: descriptive statistcs

2 types of statistics we run in our analyses

• descriptive statistics

  • typically univariate (1 variable at a time)

• relational statistics

  • inferential

  • 2 or more variables

  • focuses on the independent and dependent

descriptive statistics

• statistics that are typically summarize the data on 1 variable

• can also summarize the relationship between 2 variables w/o any hypothesis testing (bivariate or crosstabulation)

what do we use to summarize data (1)

• measures of distribution

  • frequency

    • number of participants who have a certain characteristic

  • proportion

    • frequency of participants w/ certain characteristics (n) divided by the total number of participants (n) surveyed

    • ranges from 0 to 1, typically reported as decimal

  • percent

what do we use to summarize data??

• measure of distribution

• measures of central tendency

• measures of central tendency

• measure of dispersion or variability

what do we use to summarize data (2)

• measures of central tendency

  • mode

    • most frequent value

    • rarely ever reported

  • mean

    • only report the mean when it is being normally distributed

    • report for individual or composite (scored) variables

  • median

    • report for skewed score

    • ordinal or skewed variable

what do we use to summarize data (3)

• measures dispersion or variability

  • standard deviation (s or SD)

    • extent to which participant score differ from the mean

    • reported w/ the mean (measure of central tendency)

    • like mean, also sensitive to extreme scores

  • quartiles

    • median aka 2nd quartile

  • range

    • difference between the highest & lowest score

    • report lower & upper limits

    • reported for continuous variables

  • interquartile range (IQR)

    • differece b/w the 75th & 25th percentiles

    • range provided for median

    • IQR = Q3 - Q1

scoring

working w/ scales & analyzing data

what do we use to summarize dataaaa!!

• nominal & ordinal

  • frequency

  • percent

• continuous interval & ratio

  • individual or scored (composite variable)

  • measures of central tendency & dispersion

    • mean & standard deviation OR (no mean for age ex)

    • median & quartiles

  • answer is any number

  • depends on distribution

    • normal (mean)

    • skewed (median)

frequency distribution

• table that gives info about frequencies for each score or characteristic in variable

• gives descriptive stats

  • frequency & percent

  • measure of central tendency

  • measure of dispersion

  • missing data

how do we visualize nominal & ordinal data?

• manuscript/paper: table of results

• presentation or poster: bar graph or pie chart

  • x = specific characteristics

  • y = frequency, percent, or valid % of people w/ that specific characteristic or score

how do we visualize continuous data?

• typically, table of results

• histograms

helppp

postive skews

• scores are generally low & there are fewer high scores (forming the right tail)

  • high (left) → low (right)

negative skews

• scores are generally high & there are fewer low scores (forming the left tail)

  • low (left) → high (right)

bimodal distribution

probability distribution that has 2 distinct peaks or modes, which indicates the presence of 2 separate groups or processes within the same dataset

reporting descriptive statistics

• in method & results section

• no interpretation, instead state the facts