PY3 Research Methods (adv & disadv)

Experimental Method:

• ADV: - CONTROL OVER V'S = REDUCE CONFOUNDING V'S - RESEARCHER MORE SURE THAT IV CAUSED DV & REPLICABLE...
• DISADV: -LACK ECO VALIDITY... & CREATE DEMAND CHARACTERISTICS...

Lab Experiment:

• ADV: -MOST CONTROL MEANING IT IS POSSIBLE TO SAY IV CAUSED DV & REPLICABLE...
  
• DISADV: -LACK ECO VALIDITIY... & CREATE DEMAND CHARACTERISTICS...

Field Experiment:

• ADV: -BETTER ECO VALIDITY... & REDUCED DEMAND CHARACTERISTICS...
• DISADV: -LESS CONTROL OVER V'S... LESS CERTAIN IV CAUSED DV... & MORE DIFFICULT TO REPLICATE...

Correlation:

• ADV: - TEST WHETHER RELATIONSHIP BETWEEN CO-VARIABLES & STRENGTH/NATURE OF IT, USEFUL WHEN NOT PRACTICAL/ETHICAL TO MANIPULATE V'S
• DISADV: - NOT POSSIBLE TO EST CAUSE&EFFECT

Natural Experiment:

• ADV: -GOOD ECO VALIDITY, REDUCED DEMAND CHARACTERISTICS
  
• DISADV: - LESS CONTROL OVER V'S, LESS CERTAIN THAT... & MORE DIFFICULT TO REPLICATE

Observation:

• ADV: - A LOT OF QUAL DATA & GOOD ECO VALIDITIY 
• DISADV: - QUAL = DIFFICULT TO ANALYSE STATISTICALLY, DIFFICULT TO EST CAUSE&EFFECT AND NO CONTROL OVER V'S

Questionnaires:

Closed q's - ADV: - DATA EASIER TO ANALYSE DISADV: - FORCED ANSWERS MAY REDUCE VALIDITY & LTD INFO ABOUT REASONS BEHIND RESPONSES

Open q's - ADV: - RICHER QUALITY OF INFO (QUAL) DISADV: - DIFFICULT TO ANALYSE STATISTICALLY 

General: ADV: - LARGE AMOUNTS OF DATA EASILY & CHEAP DISADV: MAY CONTAIN LEADING Q'S OR Q'S MAY BE UNCLEAR/AMBIGUOUS & D CHARACTERISTICS E.G. SOCIAL DESIRABILITY

Interviews:

Structured - ADV: - Q'S STANDARDISED, EASIER TO SUMMARISE DATA & COMPARE DISADV: - LESS FLEXIBLE 

Unstructured - ADV: - MORE FLEXIBLE DISADV: - OPEN TO BIAS, DIFFICULT TO ANALYSE

Interviews:

General: 

• ADV: - QUALITATIVE DATA - REASONS BEHIND BEHAVIOURS
• DISADV: - QUALITATIVE DATA DIFFICULT TO ANALYSE STATISTICALLY 

Qualitative:

• ADV: -RICH IN DETAIL, MORE IN-DEPTH INFO, UNDERSTAND ATTITUDES, OPINIONS & BELIEFS
• DISADV: -HARDER TO ANALYSE STASTICALLY- MUST BE QUANTIFIED 1ST WHICH MAY RESULT IN LOSS OF DEPTH & DETAIL

Quantitative:

• ADV: -EASIER TO ANALYSE STATISTICALLY, PATTERNS CAN BE IDENTIFIED & INFERENTIAL STATS USED TO SHOW SIGNIFICANCE
  
• DISADV: -LESS DEPTH & DETAIL

Nominal level:

• ADV: -QUICK & EASY TO ANALYSE, INFERENTIAL STATS CAN BE USED TO SHOW SIGNIFCANCE
  
• DISADV: -LESS INFO THAN OTHER LEVELS E.G. INTERVAL OR RATIO

Ordinal level:

• ADV: -MORE INFO THAN NOMINAL, PPS SCORES CAN BE COMPARED & INFERENTIAL STATS CAN BE USED TO SHOW SIGNIFICANCE
• DISADV: -LESS INFO THAN INTERVAL OR RATIO - PARTICULAR PROB = DIFFERENCE BETWEEN RANKS MAY VARY CONSIDERABLY

Interval/ratio level:

• ADV: -MORE PRECISE & GIVE MORE INFO THAN NOMINAL AND ORDINAL - MEASURED IN FIXED UNITS WITH EQUA DIFF BETWEEN ALL POINTS ON THE SCALE
  
• DISADV: -INTERVAL GIVES LESS DETAIL THAN RATIO AS 0 POINT ON SCALE IS NOT ABSOLUTE E.G. SCORE OF 0 ON A STRESS SCALE DOES NOT MEAN THERE IS NO STRESS

Independent groups:

• ADV: -AVOIDS ORDER EFFECTS E.G PRACTICE ( CAN BE A PROB IN REPEATED MEASURES), REDUCES THE LIKELIHOOD OF DEMAND CHARACTERISTICS...
• DISADV: -DIFFERENCES BETWEEN GROUPS MAY BE DUE TO INDIVIDUAL DIFFERENCES RATHER THAN MANIPULATION OF THE IV

Repeated measures:

• ADV: -ANY DIFFS BETWEN CONDITIONS CANNOT BE DUE TO INDIV DIFFS ( ADV OVER INDEPENDENT GROUPS ), REQUIRES LESS PPS THAN INDEPENDENT GROUPS 
• DISADV: -ORDER EFFECTS E.G. FATIGUE & PRACTICE ( NOT IN I GROUPS ), MAY AFFECT RESULTS IN 2ND PART OF STUDY

Matched pairs:

• ADV: -OVERCOMES PROBLEM OF INDIVIDUAL DIFFERENCES, AVOIDS ORDER EFFECTS, LESS LIKELY TO GET D CHARACTERISTICS 
• DISADV: -CANNOT BE PERFECTLY MATCHED (STILL A RISK OF PP VARIABLES) , TIME CONSUMING TO MATCH

Random sampling:

• ADV: -MORE LIKELY TO BE UNBIASED, CHANCE OF SELECTING ALL OF A PARTICULAR SUB-GROUP IS MINMAL
• DISADV: -DOES NOT GUARANTEE A REPRESENTATIVE SAMPLE (CERTAIN SUB-GROUPS MAY NOT BE SELECTED), MORE TIME-CONSUMING AND DIFFICULT THAN OTHER TECHNIQUES

Opportunity:

• ADV: -QUICKER AND EASIER THAN OTHER TECHNIQUES E.G. RANDOM 
• DISADV: -EASILY BIASED, LESS LIKELY TO BE REPRESENTATIVE OF TARGET POP, CERTAIN SUB-GROUPS MAY NOT BE AVAILABLE

Volunteer/self-selecting:

• ADV: -LIKELY TO BE MOTIVATED, EASIER THAN OTHER TECHNIQUES E.G RANDOM 
• DISADV: -MAJORITY OF TARGET POP UNLIKELY TO RESPOND, THOSE WHO DO MAY NOT BE TYPICAL, POTENTIALLY BIASED & UNLIKELY TO BE REPRESENTATIVE

Systematic sampling:

• ADV: -LESS LIKELY THAT RESEARCHER WILL SELECT PPS ON PERSONAL ATTRIBUTES THAN OTHER TECHNIQUES E.G. OPPORTUNITY, EASIER THAN SOME TECHNIQUES E.G. RANDOM
• DISADV: -NOT EVERY PERSON IN SAMPLING FRAME HAS AN EQUAL CHANCE OF SELECTION, PEOPLE MAY NOT WANT TO PARTICIPATE AND THEN THE SYSTEM BREAKS DOWN

Mean (CT) :

• ADV: -USES ALL AVAILABLE DATA, MOST SENSITIVE MEASURE OF CT
• DISADV: -MAY NOT GIVE AN ACCURATE INDICATION OF CT AS IT CAN BE DISTORTED BY EXTREME VALUES

Median (CT):

• ADV: -USEFUL IF THERE ARE EXTREME VALUES, UNLIKE THE MEAN THE MEDIAN IS UNAFFECTED BY OUTLYING VALUES
• DISADV: -NOT ALL DATA USED - LESS SENSITIVE MEASURE OF CT THAN THE MEAN
  

Mode (CT):

• ADV: -ALWAYS A VALUE WHICH ACTUALLY OCCURS IN THE DATA 
• DISADV: -MAY NOT BE A SINGLE MODAL VALUE, MAY BE SEVERAL

Range (DISP):

• ADV: -QUICKER & EASIER THAN STANDARD DEVIATION
• DISADV: -UNLIKE SD, DOESNT PROVIDE INFO ABOUT DISTRIBUTION OF SCORES AROUND THE MEAN

Standard deviation (DISP):

• ADV: -ALL DATA USED - MORE SENSITIVE THAN RANGE, ALSO TELLS US ABOUT DISPERSION AROUND THE MEAN
• DISADV: -REQUIRES DATA TO FORM A NORMAL DISTRIBTION CURVE - LESS EFFECTIVE IF DISTRIBUTION OF SCORES IS SKEWED OR THERE ARE OUTLYING/EXTREME SCORES