Fichas
RDA Basics
Fichas del conjunto (14)
- 1QUANTITATIVE APPROACH *Mainstream scientific approach in psych > Empirical/experimental research > simple numerical data, measurement, explanationQUALITATIVE APPROACH
> Uses textual data: exploratory, complex rich data
> Deals with how people understand their experiences - 2BASIC RESEARCH > How it works... Whats going on > Figuring out relations: General, CasualAPPLIED RESEARCH
> Solving real-world problems
> External validity - 3TYPES OF HYPOTHESES: CORRELATIONAL RELATIONSHIP > Descriptive > 2 things perform in a synced manner X: confounding variableTYPES OF HYPOTHESES:
CASUAL RELATIONSHIP
> What is the hierarchical organisation of the variables
> Measuring how a correlation impacts a relationship - 4NULL HYPOTHESIS H0: This is set to be nullified or refuted - assumed to be true until statistical evidence indicates otherwiseALTERNATIVE HYPOTHESIS H1: Supports your prediction
- 5IV: The variable changed DV: The variable measured Control V: Kept constant? controlled Confounding V: Variable which may confound our resultsNOMINAL DATA: Name only
> Categorical (records categories)
> DISCRETE
> Not real NO.s (can't + or -) NO MEAN
> Can't be considered higher or lower
> E.g. Male or Female (M or F)
*McNEMAR TEST & CHI SQUARE TEST - 6ORDINAL: Order of Magnitude > Categorical: ranks, positions, ordered > E.g. 1st, 2nd, 3rd in a race > Not real numbers, but can be higher or lower: CAN'T USE MEAN *SPEARMAN CORRELATION, MANN-WHITNEY TEST & WILCOXON TESTSCALE: EQUIDISTANT UNITS
> CONTINUOUS or cardinal
> Much more useful
> Real numbers: can be averaged & USE MEAN and others
> E.g. cm, kg, ages
*PEARSON CORRELATION PAIRED T-TEST & INDEPENDENT T-TEST - 7BETWEEN DESIGN: Different participants in each condition > Independent/ Unrelated/ Unpaired/ Uncorrelated WITHIN DESIGN: Same participants in each condition > Repeated/ Related/ Paired/ CorrelatedNATURALISTIC OBSERVATIONS
> Non-numerical or numerical measure
- Qualitative or Quantitative
*High level of realism - 8PSYCHOMETRICS: Quantitative science of measuring psychological constructs *Statistics play a major role - Questionnaires, General Ability Measures: IQ tests, creativity testsSYSTEMATIC OBSERVATION
> Experimenter has a great control over a variable of a study: gives them insight as to how the variables are related
- Good Internal validity
- External Validity issue - 9CORRELATIONAL STUDIES > Relationship between variables - 2 DVs: predictor & outcome > Not casual: Correlation doesn't mean causation > Relationship Types: Negative, positive, no relationship > Scatter plotsEXPERIMENTAL METHOD
> 2 identical situations: difference of outcomes caused by one differing variable
- Change in variable X (comes before) associated with change in Y
- Has confounding variables - 10TRUE EXPERIMENT: investigation of casual relationships between variables (because of randomisation): Random sample > vary an IV and measure impact on the DV Y: results are probabilistic Y: several variations: between, within, mixed X: V narrowly defined = lack construct validity X: lacks ecological validity X: can produce artificial findingsQUASI-EXPERIMENT: assignment is pre-determined
- Applied research where true experimental design cannot be achieved
- QUANTITATIVE METHOD - 11MEASURES OF CENTRAL TENDENCY MODE: most frequent value in a dataset: SCALE, ORDINAL, NOMINAL MEDIAN: value in the middle of a ranked dataset: SCALE, ORDINAL MEAN: the average: SCALEMEASURES OF DISPERSION
RANGE: the difference between highest & lowest scores
IQ: (preferred): considers the central 50% of a distribution. *BPX AND WHISKER (BOXPLOT)
VARIANCE: MEAN, MEDIAN: considers how much each score differs from the mean
SD - 12CONVERTING VARIANCE INTO SD Instead Range/IQR, better consider all scores, we: Measure deviation of each score from mean: (x - ẋ) Square these to get rid of +/- problem: (x - ẋ)2 Add these to get figure for all sample: Σ (x - ẋ)2 Divide by n-1 to get average for sample: Σ (x - ẋ)2 /n-1 Take square root of variance to get StDev: √ [Σ (x - ẋ)2 /n-1]CENTRAL TENDENCY & LOM
NOMINAL: mode
ORDINAL: median (+mode)
SCALE: mean (median+mode)
*mean is preferred as it allows parametric testing - 13DISPERSION AND LOM NOMINAL: N/A ORDINAL: Range or IQR SCALE: St.d or variance (*St.d preferred as root returns to original level)Z SCORES: the number of SDs from the mean (preferred as takes mean as its focus)
T SCORES: Z score transformed to be always +ve value - 14—