Docstoc

Basic Concepts

Document Sample
Basic Concepts Powered By Docstoc
					Basic Concepts
  a brief review
Research Studies
      Non-Experimental                   Experimental
        (Descriptive)                    (Laboratory)
   Behavior is observed as it      Manipulate one variable to
    naturally occurs                 see the effect on another
                                     variable (cause-and-effect
                                     relationship)
   Advantage: natural
                                    Advantage: control
                                     (representative sampling,
   Disadvantage: lack of            random assignment,
    control                          manipulation)
                                    Disadvantage: artificial;
                                     ethical limitations
Variables
    Independent              Dependent

      The variable            The variable
  manipulated by the        measured by the
  researcher and is the   researcher and is the
     “cause” of the          “effect” of the
        behavior          independent variable
Variables
   The independent variable must have at
    least two conditions or levels

      age, gender, intelligence, aggression, anxiety

                 variable: gender
                    levels: ???
What do you think?
  Gender differences in voting preference
              (Bush or Kerry)



    Variable =>        Levels =>

    Variable =>        Levels =>
Experimental Studies…
           Does driving a SUV cause people
             to drive more aggressively?

   Independent Variable     Dependent Variable

       type of car                 aggression
      (sedan, SUV)
Experimental Studies…
   Control group: The baseline or standard
    condition  sedan

   Experimental group: Receives some level of the
    independent variable  SUV

   If behavioral changes occur when the
    independent variable is manipulated then we can
    conclude… the independent variable caused
    changes in the dependent variable
Between-Participants
   The participants in each group are
    different
   Different people serve in the control and
    experimental groups
       Random sample
       Random assignment
Within-Participants (Repeated-measures)
   The same participants are used in all
    conditions
   Repeatedly taking measures on the same
    individuals
   The same people serve in the control and
    experimental groups
       Random sample
       NOT random assignment
For example…
   How would you conduct a study in this
    class on the effects of a mnemonic device
    on memory?
       Between-Participants

       Within-Participants
Within vs. Between
     Within-Participants              Between-Participants

   Advantage: Requires less         Advantage: Minimizes
    participants and time,            order, carryover effects,
    increased statistical power       and demand characteristics
    (less variability due to
    individual differences)

   Weakness: Order effects          Weakness: Requires more
    (practice and fatigue            participants and time,
    counterbalance), carryover        groups may not be equal,
    effects, demand                   less powerful statistically
    characteristics
Statistics
a brief review
Statistics: What’s the point?
   Observations researchers make are DATA
   STATISTICS are a set of mathematical
    procedures for summarizing and
    interpreting data
Statistics: What’s the point?
   Two types of statistics:
       Descriptive Statistics (“the easy ones”):
        Summarize the data into one or two
        representative numbers

        Inferential Statistics: Make inferences
        about the meaning of the data (infer that the
        results from the sample apply to the
        population)
Descriptive Statistics
       Measures of Central Tendency
       (characterizes the typical response)

   Mean: The average of a group of scores
   Median: The middle score in a
    distribution
   Mode: The most frequent score
Descriptive Statistics
       Group A           Group B
        8                 27
        6                 6
        5                 5
        5                 5
        4                 4
        3                 3
        3                 3
        3                 3
        1                 1

       Mean = 4.2        Mean = 6.3
       Median = ???      Median = ???
       Mode = ???        Mode = ???
Descriptive Statistics
                 Variability
       (The amount by which subjects
           vary from one another)

            Standard Deviation
 (The “standard” or normal amount the scores
     deviate or move away from the mean)
Descriptive Statistics

    SD =  (x – M)2
              N


  x = each score
  M = mean
  N = total number of scores
For Example…
                  Scores
                2, 2, 4 , 4

 Mean= ???
= 2 + 2 + 4 + 4 / N
 = 12 / 4
= 3
For Example…
                           Scores
                         2, 2, 4 , 4
   SD= ???

   SD =  (x – M )2
            N

   SD = (2 - 3)2 + (2 - 3)2 + (4 - 3)2 + (4 - 3)2 / N


   SD = 1 + 1 + 1 + 1 / 4

   SD = 1
For Example…
   Group A: 0, 5, 10, 15, 20, 25, 30
      Mean = 15.00 SD = 10.00

  Group B: 14, 14, 14, 15, 16,16, 16
       Mean = 15.00 SD = .93

  Group C: 15, 15, 15, 15, 15, 15, 15
       Mean = 15.00 SD = ???
Descriptive Statistics in a Nutshell
   Central tendency tells you about the
    “average” person

   Variability tells you how much people
    differ from the “average” person

   Great for organizing and summarizing
    data but they are only 1/2 the picture…
 How do you deicide if group
differences are reliable and not
      due to chance alone?
Inferential Statistics
   It is impossible to study the entire population
    therefore we must use a sample of the population

   Inferences are made about the likelihood that the
    differences in a sample reflect true differences in
    the population

   Inferential statistics tell us if differences in the
    sample are large enough to conclude that there
    are differences in the population
    (“moving from the sample to the population”)
Inferential Statistics
   Data always consist of 2 components:

       The participant’s actual value on the dimension
        being measured


       Errors of measurement
Inferential Statistics
   Inferential statistics establish a probability
    that the results are real (not due to error)

   Data with a probability (alpha level) less
    than 5% (p = .05) are regarded as
    statistically significant

        This means  5 times out of 100 the results
        are going to be due to random error
Points to Ponder
   Statistical significance allows us to say
    that our results are probably not due to
    chance

   ‘Significance’ only refers to statistical
    probability, not to the theoretical or
    practical importance
Levels of Measurement
  Lowest     Scale      Characteristic        Example          Limitations
   Level
                                                              Categories are
            Nominal         Naming          Males; Females   not in any order
                         (no numbers)                             (Mode)
                                            Ranking people    Not able to look
            Ordinal        Ordering           by level of    at the size of the
                                             aggression         differences
                                                                   (Mdn)
                        Equal intervals                           Not make
            Interval   without absolute     Thermometer         proportional
                             zero                               statements
                                                                  (M, SD)
                         Equal intervals
             Ratio     with absolute zero   Weight; Height        None
                                                                 (M, SD)
  Highest
   Level
Choosing a Statistic
                     My research question
                          is about …



    Differences           Frequencies         Relationships
  between means                             between variables


   t test, ANOVA          Chi-Square          Correlation
                                              Coefficients


  Interval / Ratio         Nominal               Ordinal
                                              Interval/Ratio

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:2
posted:8/15/2012
language:
pages:30