Basic Statistics Vernon E. Reyes
Topics What is statistics in relation to social science research? Organizing data Measures of central tendency Measures of variability Probability, samples and population Testing differences between means ANOVA Chi-Square Correlation Spearmans ro
Social Science Research Social science = study of the obvious? Thus there is a need to test hypotheses Example: Mass Murderers *which of the following are obvious and therefore not worth studying?
Mass murderers? Mass murderers are almost ALWAYS insane (a person who is sane will never kill people who are having lunch) Mass murderers are usually strangers to their victims who are unlucky enough to be at the wrong place and at the wrong time Mass murderers typically run amok, expressing themselves in a spontaneous outpouring of anger Mass murderers look different from the rest of us
However… 697 mass killers from 1976 – 1995 Mass murderers are rarely insane, they know what they are doing and NOT because “voices of demons” Random shootings in public places are rare, most killings happen within families or acquiantances Most murderers are methodical and selctive NOT impulsive or spontaneous Murderers do not look different, they look just ordinary people
Social science research Social scientists attempt to explain and predict human behavior In the process, soc. Scientists examine characteristics of human behaviors called variables – characteristics that differ from one individual to another (i.e. age, social class, attitude) or from one point in time to another (i.e. unemployment, crime, population). Constant: gender of mothers Variable: age, race, mental health of mothers
Are there different variables? Independent variable/s (IV) - This is the one presumed to cause the changes in the DV Dependent variable/s (DV) This is one presumed to to change or be the effect because of your IV. Give your own examples…
Series of numbers for social science research Series of numbers have a level of measurement. 1. Classify or categorize = nominal level of measurement 2. Rank or order = ordinal levels of measurement 3. Assign a score = interval level (no true value of zero, i.e. temperature) or ratio (has an absolute zero i.e. weight, height, etc) of measurement
NOMINAL LEVEL Involves naming or labeling, meaning we place cases into categories and count frequency Examples: gender (male, female), political party (Lakas, Liberal, Nacionalista), time (past, present, future) etc
Notes for nominal data Every case must be placed in one, and only 1, category = categories must not be overlapping or mutually exclusive Must be exhaustive = a place for every case that arises Nominal data are not graded, ranked or scaled… i.e. better or worse, higher or lower, more or less
Ordinal Level Degree of ranking Example: instead of just categorizing nationalists versus not nationalist, we might want to study the degree of nationalism Not a measure of magnitude. i.e. we do not know how nationalist one respondent from another. Example, beauty contests
Interval/Ratio level Not only shows order but also exact distance of one score to another. Uses constant units of measurements (pesos, celsius, yards, feet, minutes) and yield equal intervals
What is what?! There is clear distinction between nominal data and ordinal data Example: color of hair [black, brown, blonde] (nominal) vs condition of hair [dry, normal, oily] (ordinal) Distinction between ordinal and interval is not always clear cut
Attitude toward professor Scale value Rank of professor Attitude toward professor 1 University professor Very favorable 2 Full professor Favorable 3 Assoc. professor Somewhat favorable 4 Asst. Professor Neutral 5 Instructor Somewhat unfavorable 6 Lecturer Unfavorable 7 Teaching asst. Very unfavorable
From the previous table Difference between instructor and lecturer is minimal in terms of prestige, salary or qualifications The difference between instructor and asst. professor is substantial wherein there should be at least a doctorate and higher salary Attitude is evenly spaced ie strongly agree – strongly diagree instead, whenever possible, we treat ordinal variables as interval ONLY if safe to assume that the scale has relatively equal intervals! Attitude = interval and Rank of Prof = ordinal only!
Functions of Statistics Description Decision making / inferences