Notesale: Turn your study into money

Document Preview

Extracts from the notes are below, to see the PDF you'll receive please use the links above

---

3:​ ​LIBRARY​ ​RESEARCH

1

5&6:​ ​STATISTICS

2

7:​ ​ACADEMIC​ ​MISCONDUCT​ ​AND​ ​PLAGIARISM

7

8:​ ​POSTER​ ​TIPS

7

9:​ ​USING​ ​EXCEL

8

10:​ ​CRITICAL​ ​READING​ ​AND​ ​WRITING

8

11:​ ​REFERENCING

11

12&14:​ ​DESIGNING​ ​AN​ ​INVESTIGATION

13

13&16:​ ​PRESENTING​ ​DATA

14

15:​ ​TIME​ ​MANAGEMENT

17

17:​ ​REVISION​ ​AND​ ​EXAMS

17

Joanna​ ​Griffith​ ​(2017)

3:​ ​LIBRARY​ ​RESEARCH
Extra​ ​information​ ​(outside​ ​lectures)​ ​is​ ​needed​ ​for:
○ Background​ ​reading​ ​around​ ​a​ ​research​ ​topic
○ Comparisons​ ​with​ ​primary​ ​research​ ​from​ ​articles​ ​in​ ​scientific​ ​journals
● Hierarchy​ ​of​ ​resources:
○ Original​ ​articles​ ​(containing​ ​primary​ ​research)
○ Literature​ ​reviews​ ​(secondary​ ​research)
○ Books
○ Organisational​ ​websites​ ​(for​ ​government​ ​and​ ​scientific​ ​organisations)
○ blogs/Wikipedia/other​ ​non-official​ ​websites
The​ ​Penryn​ ​library
● Loans:
○ Self-issue​ ​and​ ​return​ ​24/7
○ Allowance​ ​of​ ​20​ ​items​ ​and​ ​2​ ​DVDs
○ Loan​ ​period​ ​of​ ​6​ ​weeks,​ ​unless​ ​books​ ​are​ ​High​ ​Demand​ ​or​ ​Reference​ ​copies
● Finding​ ​books​ ​for​ ​biosciences:
○ Biodiversity​ ​=​ ​333
...
00
○ Genetics​ ​=​ ​576
...
80
○ Ecology​ ​=​ ​577
...
00
○ Animal​ ​behaviour​ ​=​ ​591
...
00
○ Birds​ ​=​ ​598
...
00
● Catalogue​ ​=​ ​lib
...
ac
...
​ ​names/numbers)
○ Ordinal
■ Orders​ ​observations​ ​along​ ​a​ ​continuum​ ​(eg
...

concentration,​ ​weight,​ ​height)
Central​ ​tendency
● Measures​ ​the​ ​spread​ ​of​ ​data
● If​ ​you​ ​take​ ​repeated​ ​samples​ ​from​ ​a​ ​population​ ​and​ ​calculate​ ​their​ ​averages​ ​they​ ​will
be​ ​normally​ ​distributed​ ​(Central​ ​Limit​ ​Theorem)
● Central​ ​tendency​ ​can​ ​be​ ​described​ ​using​ ​mean,​ ​median,​ ​and​ ​mode
○ Don’t​ ​use​ ​all​ ​of​ ​them​ ​at​ ​once,​ ​think​ ​about​ ​the​ ​characteristics​ ​of​ ​the​ ​data
● Mean
○ Where​ ​do​ ​most​ ​of​ ​the​ ​observations​ ​lie?
○ Always​ ​an​ ​estimation​ ​of​ ​the​ ​population​ ​mean,​ ​and​ ​is​ ​referred​ ​to​ ​as​ ​the​ ​sample
mean
○ Sample​ ​mean​ ​=​ ​sum​ ​of​ ​each​ ​observation/number​ ​of​ ​observations​ ​in​ ​the
sample
○ Problems:
●

Joanna​ ​Griffith​ ​(2017)

■
■

○

Skewed​ ​data​ ​strongly​ ​distorts​ ​the​ ​mean
Does​ ​not​ ​work​ ​as​ ​well​ ​as​ ​other​ ​central​ ​tendency​ ​techniques​ ​if
exceptional​ ​values​ ​in​ ​the​ ​data​ ​distort​ ​clustering,​ ​or​ ​if​ ​data​ ​has​ ​a​ ​trend
(scatter​ ​plots),​ ​or​ ​is​ ​bimodal

Rules:
■ The​ ​mean​ ​of​ ​a​ ​set​ ​of​ ​observations:
● Does​ ​not​ ​indicate​ ​that​ ​any​ ​of​ ​the​ ​observations​ ​have​ ​that​ ​exact
value
● Some​ ​of​ ​the​ ​values​ ​in​ ​the​ ​dataset​ ​may​ ​not​ ​even​ ​be​ ​close​ ​to​ ​the
mean​ ​value
■ Means​ ​can​ ​be​ ​used​ ​safely​ ​when​ ​observations​ ​cluster​ ​around​ ​a​ ​central
value​ ​(central​ ​tendency)
● Median
○ Middle​ ​observation​ ​in​ ​a​ ​dataset​ ​where​ ​values​ ​have​ ​been​ ​ranked​ ​in​ ​terms​ ​of
magnitude
○ Rank​ ​data​ ​in​ ​increasing​ ​order,​ ​and​ ​find​ ​the​ ​middle​ ​value
■ If​ ​there​ ​are​ ​even​ ​numbers​ ​of​ ​observations​ ​in​ ​the​ ​set,​ ​use​ ​the​ ​mean​ ​of
the​ ​two​ ​middle​ ​observations
○ Rules:
■ ‘Resistant​ ​statistic’​ ​that​ ​is​ ​less​ ​susceptible​ ​to​ ​outliers​ ​or​ ​extreme
values,​ ​unlike​ ​mean
● Better​ ​for​ ​use​ ​with​ ​skewed​ ​datasets
■ More​ ​affected​ ​by​ ​changes​ ​in​ ​values​ ​near​ ​the​ ​middle​ ​of​ ​the​ ​distribution
■ Not​ ​easily​ ​written​ ​as​ ​an​ ​equation
■ Hard​ ​to​ ​use​ ​with​ ​large​ ​datasets
● Mode
○ The​ ​value​ ​in​ ​any​ ​set​ ​of​ ​observations​ ​which​ ​occurs​ ​most​ ​frequently
○ Rules:
■ Can​ ​indicate​ ​a​ ​“normal”​ ​or​ ​“usual”​ ​figure
■ Use​ ​when​ ​numbers​ ​in​ ​a​ ​distribution​ ​are​ ​not​ ​evenly​ ​distributed​ ​around
a​ ​central​ ​value
■ Is​ ​a​ ​value​ ​that​ ​actually​ ​occurs​ ​(unlike​ ​median,​ ​which​ ​can​ ​be​ ​a​ ​decimal)
■ Not​ ​useful​ ​if​ ​there​ ​is​ ​no​ ​single​ ​most​ ​frequently​ ​occurring​ ​value
Descriptive​ ​statistics
● Visualise​ ​the​ ​spread/distribution​ ​of​ ​data​ ​values
○ Eg
...
​ ​between​ ​different​ ​species
Skewness​ ​and​ ​kurtosis
● Skewness:​ ​degree​ ​of​ ​displacement​ ​in​ ​the​ ​peak
○ The​ ​closer​ ​the​ ​peak​ ​is​ ​to​ ​the​ ​middle,​ ​the​ ​less​ ​skewed​ ​it​ ​is
● Kurtosis:​ ​‘peakedness’​ ​(degree​ ​to​ ​which​ ​data​ ​values​ ​are​ ​concentrated​ ​at​ ​one​ ​point)
Summary​ ​of​ ​descriptive​ ​statistics
● Descriptive​ ​statistics​ ​are​ ​used​ ​to​ ​condense​ ​data​ ​into​ ​a​ ​single​ ​meaningful​ ​value
(mean,​ ​median,​ ​and​ ​mode)​ ​and​ ​illustrate​ ​central​ ​tendency
○ Don’t​ ​use​ ​all​ ​of​ ​them​ ​at​ ​once,​ ​consider​ ​distribution​ ​of​ ​data,​ ​skewness,​ ​and
kurtosis
Joanna​ ​Griffith​ ​(2017)

Standard​ ​deviations​ ​are​ ​the​ ​building​ ​blocks​ ​of​ ​probability​ ​testing​ ​and​ ​confidence​ ​in
the​ ​dataset
○ Don’t​ ​use​ ​all​ ​of​ ​them​ ​at​ ​once,​ ​consider​ ​distribution​ ​of​ ​data,​ ​skewness,​ ​and
kurtosis
● Standard​ ​deviations​ ​are​ ​the​ ​building​ ​blocks​ ​of​ ​probability​ ​testing​ ​and​ ​confidence​ ​in
the​ ​dataset
Statistical​ ​tests​ ​for​ ​significance
● Parametric​ ​tests
○ Make​ ​strict​ ​assumptions​ ​about​ ​the​ ​data​ ​being​ ​analysed
○ Requires​ ​data​ ​to​ ​be​ ​normally​ ​distributed
○ Requires​ ​around​ ​50+​ ​observations
● Non-parametric​ ​tests
○ Not​ ​based​ ​upon​ ​stringent​ ​assumptions​ ​of​ ​normality
■ Therefore​ ​slightly​ ​weaker
○ Can​ ​be​ ​run​ ​on​ ​small​ ​data​ ​sets​ ​(<20)
● Statistical​ ​tests​ ​for​ ​level​ ​1​ ​(accessed​ ​through​ ​Excel)
○ To​ ​use​ ​Excel​ ​to​ ​find​ ​the​ ​difference​ ​between​ ​data​ ​on​ ​two​ ​samples,​ ​you​ ​must
have:
■ Two​ ​measured​ ​variables​ ​or​ ​two​ ​counts​ ​of​ ​things
■ More​ ​than​ ​15​ ​observations​ ​per​ ​sample
■ An​ ​equal​ ​number​ ​of​ ​observations​ ​for​ ​each​ ​variable
○ Tests​ ​for​ ​a​ ​difference
■ Student’s​ ​t-test
● Sample​ ​against​ ​population
■ Difference​ ​of​ ​means​ ​t-test
● Two​ ​individual​ ​samples
■ T-test​ ​for​ ​tied​ ​samples
● Measures​ ​of​ ​the​ ​same​ ​set​ ​of​ ​organisms​ ​at​ ​two​ ​time​ ​points
○ Tests​ ​for​ ​a​ ​relationship​ ​(correlation)
■ Are​ ​the​ ​two​ ​variables​ ​being​ ​tested​ ​related​ ​in​ ​some​ ​way?
● Often,​ ​there​ ​is​ ​not​ ​a​ ​perfect​ ​relationship​ ​because​ ​other
variables​ ​may​ ​have​ ​had​ ​an​ ​effect​ ​on​ ​the​ ​dependent​ ​variable
● The​ ​variability​ ​around​ ​the​ ​line​ ​of​ ​each​ ​point​ ​determines​ ​the
strength​ ​of​ ​the​ ​correlation
● Normality
○ A​ ​bell​ ​curve​ ​for​ ​a​ ​single​ ​sample
○ A​ ​straight​ ​line​ ​for​ ​trends/correlations
● Variance
○ Before​ ​using​ ​a​ ​statistical​ ​test,​ ​check​ ​that​ ​the​ ​data​ ​sets​ ​look​ ​different​ ​and
conform​ ​to​ ​the​ ​requirements​ ​of​ ​normality
○ Variance​ ​=​ ​standard​ ​deviation​2
○ In​ ​order​ ​to​ ​check​ ​that​ ​the​ ​data​ ​adheres​ ​to​ ​the​ ​assumption​ ​of​ ​equal​ ​variance,
an​ ​F-test,​ ​Kolmogorov​ ​or​ ​Shapiro​ ​test​ ​can​ ​be​ ​used
● Significance
●

Joanna​ ​Griffith​ ​(2017)

The​ ​means​ ​of​ ​two​ ​samples​ ​may​ ​look​ ​different​ ​from​ ​each​ ​other,​ ​but​ ​the
distributions​ ​of​ ​the​ ​measurements​ ​which​ ​were​ ​used​ ​to​ ​calculate​ ​the​ ​means
might​ ​overlap​ ​or​ ​intersect
○ The​ ​P​ ​value
■ Used​ ​to​ ​demonstrate​ ​the​ ​confidence​ ​we​ ​have​ ​that​ ​out​ ​data​ ​sets​ ​are
different​ ​from​ ​one​ ​another​ ​or​ ​correlated
■ If​ ​the​ ​value​ ​is​ ​0
...
05,​ ​n​ ​=​ ​…’
○ Correlation:
■ ‘There​ ​was​ ​a​ ​significant​ ​correlation​ ​between​ ​x​ ​and​ ​y,​ ​that​ ​was​ ​strongly
positively​ ​correlated​ ​(P-value​ ​=​ ​<0
...
05​ ​does​ ​not​ ​mean​ ​there​ ​was​ ​no​ ​underlying​ ​effect,
just​ ​that​ ​none​ ​was​ ​detected
● Extrapolating​ ​beyond​ ​the​ ​data​ ​is​ ​risky​ ​(patterns​ ​found​ ​within​ ​a​ ​data​ ​range​ ​will​ ​not
necessarily​ ​be​ ​true​ ​outside​ ​of​ ​it)
● Data​ ​can​ ​be​ ​cherry-picked,​ ​but​ ​selecting​ ​data​ ​in​ ​a​ ​huge​ ​dataset​ ​that​ ​looks​ ​like​ ​it​ ​has
a​ ​pattern​ ​doesn’t​ ​prove​ ​anything
● Pseudoreplication
● Randomisation​ ​(eg
...
​ ​by​ ​referencing
badly)
Detection
● Turnitin​ ​has​ ​a​ ​plagiarism​ ​checker
○ Checks​ ​for​ ​similarities​ ​with​ ​websites​ ​and​ ​other​ ​submitted​ ​work
○ Can​ ​detect​ ​when​ ​whole​ ​passages​ ​have​ ​been​ ​copied​ ​with​ ​words​ ​changed
○ Technical​ ​names​ ​cannot​ ​be​ ​written​ ​any​ ​other​ ​way,​ ​so​ ​will​ ​be​ ​picked​ ​up,​ ​but
markers​ ​understand
Avoiding​ ​plagiarism
● Write​ ​down​ ​details​ ​needed​ ​for​ ​referencing
● Use​ ​Harvard​ ​(Exeter)​ ​or​ ​Vancouver​ ​to​ ​reference
● Instead​ ​of​ ​copying​ ​a​ ​text,​ ​read​ ​a​ ​page​ ​and​ ​summarise​ ​it​ ​in​ ​your​ ​own​ ​words
● If​ ​you’re​ ​using​ ​an​ ​exact​ ​piece​ ​of​ ​text,​ ​use​ ​quotation​ ​marks
○ Cite​ ​author’s​ ​name​ ​and​ ​date​ ​of​ ​publication,​ ​and​ ​reference​ ​at​ ​the​ ​end​ ​of​ ​the
document
● Paraphrase
● Cite​ ​AND​ ​reference
○ It​ ​is​ ​not​ ​necessary​ ​to​ ​do​ ​this​ ​for​ ​common​ ​knowledge
--------------------------------------------------------------------------------------------------------------------------●

8:​ ​POSTER​ ​TIPS
Posters​ ​are​ ​still​ ​a​ ​major​ ​form​ ​of​ ​communication​ ​in​ ​science
○ Peer​ ​review​ ​publications
○ Scientific​ ​conference​ ​presentation
○ Scientific​ ​conference​ ​posters
Content
● Title
○ Should​ ​be​ ​visible​ ​from​ ​a​ ​few​ ​metres​ ​away
○ simple/clear
○ Eye​ ​catching​ ​but​ ​not​ ​distracting
○ Name,​ ​institution,​ ​and​ ​contact​ ​details​ ​underneath
● introduction/background
○ Set​ ​the​ ​scene
○ State​ ​why​ ​the​ ​question​ ​being​ ​investigated​ ​is​ ​important
○ Refer​ ​to​ ​(and​ ​reference)​ ​a​ ​few​ ​relevant​ ​studies
○ Bullet​ ​points​ ​are​ ​easier​ ​to​ ​read
● Methods
○ Simple,​ ​not​ ​necessarily​ ​detailed​ ​enough​ ​to​ ​be​ ​replicable
○ Give​ ​key​ ​information​ ​such​ ​as​ ​date/time
○ Details​ ​of​ ​how​ ​data​ ​was​ ​collected​ ​and​ ​prepared
○ Details​ ​of​ ​how​ ​bias​ ​was​ ​reduced
● Results
○ Don’t​ ​show​ ​raw​ ​data
○ SLAPU​ ​graphs
○ Talk​ ​about​ ​key​ ​trends​ ​(don’t​ ​interpret!)
●

Joanna​ ​Griffith​ ​(2017)

●

discussions/conclusions
○ Summarise​ ​and​ ​interpret​ ​main​ ​points
○ Compare​ ​to​ ​other​ ​published​ ​work​ ​(reference)
○ Suggest​ ​possible​ ​future​ ​work​ ​(evaluation)
■ Discuss​ ​the​ ​limitations​ ​of​ ​your​ ​own​ ​work,​ ​but​ ​don’t​ ​obsess​ ​over​ ​them
● References
○ Harvard​ ​or​ ​Vancouver
○ 3​ ​or​ ​4​ ​minimum
Making​ ​the​ ​poster
● Data​ ​collection
○ Use​ ​Excel​ ​to​ ​collect​ ​data​ ​and​ ​produce​ ​graphs
● Draft​ ​writing
● Tends​ ​to​ ​look​ ​better​ ​landscape
● Consider​ ​symmetry​ ​in​ ​layout
● Consider​ ​use​ ​of​ ​colour
● Consider​ ​use​ ​of​ ​fonts
○ San​ ​serif​ ​easier​ ​to​ ​read​ ​when​ ​enlarged
--------------------------------------------------------------------------------------------------------------------------9:​ ​USING​ ​EXCEL
--------------------------------------------------------------------------------------------------------------------------10:​ ​CRITICAL​ ​READING​ ​AND​ ​WRITING
●
●
●

●

Non-critical:​ ​addresses​ ​text​ ​at​ ​face​ ​value
Critical:​ ​addresses​ ​text​ ​as​ ​one​ ​portrayal​ ​of​ ​fact
○ Looks​ ​beyond​ ​what​ ​a​ ​text​ ​says​ ​to​ ​consider​ ​how​ ​it​ ​portrays​ ​the​ ​subject​ ​matter
How​ ​to​ ​be​ ​critical:
○ Gather​ ​complete​ ​information
○ Understand​ ​and​ ​define​ ​all​ ​terms
○ Question​ ​the​ ​methods​ ​by​ ​which​ ​facts​ ​are​ ​derived
○ Question​ ​the​ ​conclusions
○ Look​ ​for​ ​hidden​ ​assumptions​ ​and​ ​biases
○ Question​ ​the​ ​source​ ​of​ ​facts
○ Don’t​ ​expect​ ​all​ ​of​ ​the​ ​answers
○ Examine​ ​the​ ​big​ ​picture
○ Examine​ ​multiple​ ​cause​ ​and​ ​effect
○ Watch​ ​for​ ​thought-stoppers
○ Understand​ ​your​ ​own​ ​biases​ ​and​ ​values
The​ ​world​ ​is​ ​full​ ​of​ ​misleading​ ​information
○ News​ ​is​ ​exaggerated​ ​to​ ​provide​ ​a​ ​good​ ​story
○ Advertisements
○ Even​ ​academics​ ​and​ ​researchers​ ​may​ ​bend​ ​the​ ​truth,​ ​eg
...
​ ​title​ ​in​ ​italics
...
​ ​place​ ​of​ ​publication:​ ​publisher
● Chapters​ ​in​ ​books
○ Chapter​ ​author(s),​ ​publication​ ​date
...
​ ​book​ ​editor(s)
...
​ ​place​ ​of​ ​publication:​ ​publisher,​ ​page​ ​numbers
● Journal​ ​articles
○ Article​ ​authors,​ ​publication​ ​date
...
​ ​journal​ ​title​ ​in​ ​italics,​ ​volume
number​ ​(issue​ ​number),​ ​page​ ​numbers
● Internet​ ​documents
○ Document​ ​author(s),​ ​date
...
​ ​place​ ​of​ ​publication:
publisher
...
)​ ​or​ ​exploratory​ ​(already​ ​have​ ​a​ ​hypothesis​ ​in​ ​mind,​ ​making
observations​ ​prior​ ​to​ ​testing​ ​it)
Hypothesis
● A​ ​clear​ ​statement​ ​articulating​ ​a​ ​plausible​ ​explanation​ ​for​ ​observations
○ What​ ​you​ ​think​ ​is​ ​happening,​ ​based​ ​on​ ​your​ ​knowledge​ ​of​ ​theory​ ​and​ ​biology
● H​1​:​ ​alternative​ ​hypothesis
○ The​ ​behaviour​ ​that​ ​we​ ​are​ ​seeing​ ​is​ ​biologically​ ​interesting
● H​0​:​ ​null​ ​hypothesis
○ Can​ ​be​ ​statistically​ ​tested
○ ‘Nothing​ ​is​ ​going​ ​on’/conservative
● For​ ​every​ ​hypothesis​ ​that​ ​suggests​ ​something​ ​interesting​ ​is​ ​happening,​ ​there​ ​must
be​ ​a​ ​corresponding​ ​null​ ​hypothesis​ ​that​ ​states​ ​that​ ​nothing​ ​is​ ​happening
○ By​ ​testing​ ​the​ ​null​ ​hypothesis,​ ​you​ ​gain​ ​information​ ​about​ ​the​ ​alternate
hypothesis
Predictions
● What​ ​we​ ​expect​ ​to​ ​happen​ ​if​ ​our​ ​hypothesis​ ​is​ ​correct
○ Good​ ​predictions​ ​follow​ ​on​ ​logically​ ​from​ ​hypotheses​ ​and​ ​will​ ​lead​ ​to​ ​obvious
studies​ ​that​ ​allow​ ​the​ ​prediction​ ​to​ ​be​ ​tested
■ If​ ​your​ ​H​1​​ ​was​ ​correct,​ ​what​ ​would​ ​you​ ​expect​ ​your​ ​results​ ​to​ ​look​ ​like?
● Predictions​ ​about​ ​differences
○ Involve​ ​a​ ​difference​ ​between​ ​two​ ​or​ ​more​ ​groups
● Predictions​ ​about​ ​trends
○ The​ ​relationship​ ​between​ ​two​ ​continuously​ ​distributed​ ​measures
Testing​ ​a​ ​prediction
● What​ ​are​ ​you​ ​going​ ​to​ ​measure?
● How​ ​are​ ​you​ ​going​ ​to​ ​measure​ ​it?
● How​ ​many​ ​replicates​ ​do​ ​you​ ​need?
○ Replication:​ ​repetition​ ​of​ ​an​ ​experimental​ ​condition​ ​so​ ​that​ ​the​ ​variability
associated​ ​with​ ​the​ ​phenomenon​ ​can​ ​be​ ​estimated
● What​ ​variables​ ​may​ ​confound​ ​your​ ​results?
○ Assumption:​ ​a​ ​causes​ ​b​ ​or​ ​b​ ​causes​ ​a
■ In​ ​reality,​ ​c,​ ​a​ ​confounding​ ​variable,​ ​may​ ​cause​ ​a​ ​or​ ​b
●

Joanna​ ​Griffith​ ​(2017)

How​ ​will​ ​you​ ​sample​ ​your​ ​population?
○ Random
○ Systematic
○ stratified
Introduction
● Why​ ​the​ ​study​ ​was​ ​undertaken
○ State​ ​hypothesis​ ​and​ ​predictions
■ What​ ​does​ ​another​ ​person​ ​need​ ​to​ ​know​ ​about​ ​how​ ​you​ ​arrived​ ​at
these?
○ Background​ ​to​ ​the​ ​study
■ Any​ ​theoretical​ ​or​ ​previous​ ​experimental/observational​ ​work​ ​that​ ​led​ ​to
the​ ​hypothesis​ ​you​ ​are​ ​testing
■ Likely​ ​to​ ​include:
● References​ ​to​ ​previously​ ​published​ ​work
● Sometimes,​ ​a​ ​critical​ ​review​ ​of​ ​competing​ ​ideas
● A​ ​clear​ ​statement​ ​of​ ​the​ ​hypotheses​ ​and​ ​predictions​ ​to​ ​be
tested
--------------------------------------------------------------------------------------------------------------------------●

13&16:​ ​PRESENTING​ ​DATA
●

●

●

The​ ​importance​ ​of​ ​displaying​ ​data
○ Communicating​ ​information
■ Display​ ​frequencies
■ Illustrate​ ​associations​ ​between​ ​variables
■ Show​ ​differences​ ​between​ ​groups
○ Aids​ ​data​ ​analysis
■ Visualise​ ​patterns​ ​or​ ​reveal​ ​general​ ​trends
■ Important​ ​step​ ​in​ ​undertaking​ ​data​ ​analysis​ ​and​ ​applying​ ​statistical
techniques
○ Graphing​ ​should​ ​initially​ ​be​ ​explorative
■ Get​ ​a​ ​feeling​ ​for​ ​what​ ​graph​ ​would​ ​best​ ​suit​ ​the​ ​data
Guiding​ ​principles​ ​for​ ​presenting​ ​data
○ Clear,​ ​unambiguous​ ​representation​ ​of​ ​data
○ Have​ ​a​ ​point​ ​-​ ​illustrate​ ​trends​ ​and​ ​comparisons
○ Keep​ ​it​ ​simple
○ If​ ​it​ ​isn’t​ ​useful,​ ​get​ ​rid​ ​of​ ​it
Rules:
○ Appropriately​ ​descriptive​ ​axes​ ​titles,​ ​with​ ​units
○ Correct​ ​scale
○ No​ ​title,​ ​use​ ​figure​ ​number​ ​instead
■ No​ ​‘a​ ​graph​ ​to​ ​show…’,​ ​just​ ​state​ ​what​ ​the​ ​graph​ ​is
○ Big​ ​and​ ​clear
○ Use​ ​a​ ​key/legend​ ​if​ ​appropriate
○ The​ ​x-axis​ ​is​ ​always​ ​the​ ​independent​ ​variable,​ ​the​ ​y-axis​ ​is​ ​always​ ​the
dependent​ ​variable
Joanna​ ​Griffith​ ​(2017)

○ Don’t​ ​join​ ​the​ ​dots​ ​unless​ ​the​ ​x-axis​ ​is​ ​time
○ Choose​ ​a​ ​suitable​ ​graph​ ​style
Display​ ​types
● Bar​ ​charts
○ Used​ ​to​ ​display​ ​a​ ​variable​ ​for​ ​a​ ​set​ ​of​ ​categories
■ X-axis​ ​=​ ​grouping​ ​(categorical)​ ​variable
■ Y-axis​ ​=​ ​quantitative​ ​variable
○ Y-axis​ ​scale​ ​should​ ​include​ ​0​ ​to​ ​put​ ​the​ ​bar​ ​heights​ ​into​ ​context
○ Using​ ​Excel
■ Remove​ ​gridlines​ ​and​ ​graph​ ​borders
■ Remove​ ​legend​ ​(if​ ​unnecessary)​ ​and​ ​chart​ ​title
■ Reformat​ ​axes
■ Edit​ ​x​ ​and​ ​y​ ​axis​ ​labels​ ​and​ ​add​ ​axis​ ​titles
■ Modify​ ​colour​ ​as​ ​required
● Histograms
○ Uses​ ​height​ ​or​ ​area​ ​to​ ​display​ ​the​ ​frequency​ ​distribution​ ​of​ ​a​ ​single​ ​numerical
variable
■ X-axis​ ​=​ ​continuous​ ​quantitative​ ​variable,​ ​broken​ ​into​ ​discrete​ ​classes
called​ ​‘bins’
■ Y-axis​ ​=​ ​frequency​ ​of​ ​each​ ​‘variable​ ​class’​ ​in​ ​the​ ​dataset
○ Columns​ ​are​ ​continuous/adjacent
○ Excel​ ​is​ ​not​ ​great​ ​for​ ​generating​ ​histograms
● Box​ ​plots
○ Uses​ ​boxes​ ​and​ ​whiskers​ ​to​ ​display​ ​data​ ​distributions,​ ​usually​ ​media,​ ​IQR
and​ ​data​ ​range
■ X-axis​ ​=​ ​category/grouping
■ Y-axis​ ​=​ ​numerical​ ​variable
○ Concise​ ​method​ ​for​ ​displaying​ ​data​ ​distributions
○ Excel​ ​doesn’t​ ​have​ ​a​ ​function​ ​for​ ​producing​ ​box​ ​plots
● Line​ ​graphs
○ Uses​ ​dots​ ​connected​ ​by​ ​lines​ ​to​ ​display​ ​trends​ ​in​ ​a​ ​variable​ ​over​ ​time​ ​or​ ​any
other​ ​ordered​ ​series
○ Used​ ​when​ ​a​ ​categorical​ ​variable​ ​on​ ​the​ ​y-axis​ ​can​ ​be​ ​ordered​ ​or​ ​is
quantitative
○ X-axis​ ​is​ ​usually​ ​time,​ ​and​ ​a​ ​trend​ ​is​ ​mapped​ ​over​ ​increasing​ ​time
● Scatter​ ​plots
○ Used​ ​to​ ​present​ ​relationships​ ​between​ ​two​ ​variables
■ Each​ ​point​ ​on​ ​the​ ​graph​ ​represents​ ​a​ ​pair​ ​of​ ​observations​ ​(an​ ​x​ ​and​ ​a
y​ ​value,​ ​like​ ​coordinates)
■ X-axis​ ​=​ ​independent​ ​variable​ ​(driving​ ​the​ ​effect/unchangeable)
■ Y-axis​ ​=​ ​dependent​ ​variable​ ​(affected​ ​by​ ​the​ ​independent​ ​variable)
○ Can​ ​include​ ​a​ ​regression​ ​line​ ​(regression​ ​or​ ​smoothing​ ​function​ ​fitted​ ​to​ ​the
observations)
○ How​ ​do​ ​two​ ​variables​ ​behave​ ​relative​ ​to​ ​one​ ​another​ ​=​ ​correlation
○ Does​ ​variable​ ​1​ ​predict​ ​variable​ ​2​ ​=​ ​regression
● Pie​ ​charts
Joanna​ ​Griffith​ ​(2017)

○

Displays​ ​proportion/sizes​ ​of​ ​categories​ ​(a​ ​slice​ ​is​ ​proportional​ ​to​ ​the​ ​value​ ​of​ ​a
category)
Rare​ ​in​ ​science,​ ​doesn’t​ ​show​ ​much​ ​more​ ​than​ ​a​ ​table

○
Tables
○ Simple​ ​method​ ​for​ ​displaying​ ​frequencies​ ​or​ ​data​ ​summaries
■ Summary,​ ​contingency,​ ​frequency
Standard​ ​deviation/error
● Any​ ​time​ ​an​ ​‘average’​ ​is​ ​plotted,​ ​a​ ​measure​ ​of​ ​the​ ​variation​ ​associated​ ​with​ ​the
estimate​ ​must​ ​be​ ​stated
● Standard​ ​deviation​ ​is​ ​good,​ ​standard​ ​error​ ​is​ ​usually​ ​better​ ​(whichever​ ​is​ ​smallest​ ​is
best)
○ Standard​ ​error​ ​is​ ​a​ ​version​ ​of​ ​standard​ ​deviation,​ ​scaled​ ​to​ ​the​ ​amount​ ​of​ ​data
● Error​ ​bars
○ Gives​ ​a​ ​measure​ ​of​ ​how​ ​well​ ​a​ ​sample​ ​represents​ ​the​ ​population
■ When​ ​the​ ​sample​ ​is​ ​representative,​ ​standard​ ​error​ ​is​ ​small
■ Generally,​ ​if​ ​the​ ​error​ ​bars​ ​for​ ​your​ ​data​ ​do​ ​not​ ​overlap,​ ​you​ ​have
shown​ ​that​ ​there​ ​is​ ​a​ ​statistically​ ​significant​ ​difference
■ Symmetrical,​ ​on​ ​either​ ​side​ ​of​ ​the​ ​mean​ ​(above​ ​and​ ​below​ ​the​ ​plotted
point)
Interpreting​ ​data
● Results
○ ONLY​ ​what​ ​was​ ​found
○ Graph​ ​shows​ ​these​ ​trends…
○ No​ ​interpretation
● Discussion
○ What​ ​do​ ​your​ ​results​ ​mean?
○ accept/reject​ ​hypothesis​ ​and​ ​null​ ​hypothesis
○ Explain​ ​findings​ ​one​ ​by​ ​one
■ Do​ ​findings​ ​support​ ​and/or​ ​refute​ ​similar​ ​and/or​ ​related​ ​studies​ ​in​ ​the
field
○ Mention​ ​methodological​ ​limitations,​ ​but​ ​stay​ ​positive
○ Forecast​ ​next​ ​steps​ ​for​ ​research​ ​in​ ​the​ ​field
--------------------------------------------------------------------------------------------------------------------------●

15:​ ​TIME​ ​MANAGEMENT
Know​ ​what​ ​you​ ​have​ ​to​ ​do
● Make​ ​a​ ​complete​ ​to-do​ ​list
Prioritise​ ​tasks
● Sort​ ​tasks​ ​into:
○ Urgent​ ​and​ ​important
○ Urgent​ ​and​ ​not​ ​important
○ Not​ ​urgent​ ​and​ ​important
○ Not​ ​urgent​ ​and​ ​not​ ​important
Timetable​ ​time
● We​ ​all​ ​have​ ​the​ ​same​ ​amount​ ​of​ ​available​ ​time
Joanna​ ​Griffith​ ​(2017)

○ 2,016​ ​hours​ ​in​ ​a​ ​12-week​ ​term
● Create​ ​a​ ​timetable
○ Remember​ ​the​ ​fun​ ​stuff​ ​too
○ When​ ​do​ ​you​ ​work​ ​best?
● Break​ ​up​ ​large​ ​tasks​ ​into​ ​smaller​ ​chunks
Setting​ ​smart​ ​deadlines
● Specific
● Measurable
● Achievable
● Realistic
● timed
--------------------------------------------------------------------------------------------------------------------------17:​ ​REVISION​ ​AND​ ​EXAMS
● ​ ​Past​ ​exam​ ​papers:​ ​lib
...
ac
...
5​ ​hours​ ​before)
○ Avoid​ ​unnecessary​ ​distractions
○

Joanna​ ​Griffith​ ​(2017)

○ Stay​ ​relaxed​ ​but​ ​focuses
● The​ ​exam
○ Check​ ​the​ ​number​ ​of​ ​questions
○ Check​ ​both​ ​sides​ ​of​ ​the​ ​paper
○ Plan​ ​or​ ​make​ ​notes​ ​before​ ​answering​ ​if​ ​necessary
○ Keep​ ​an​ ​eye​ ​on​ ​the​ ​time
○ Allow​ ​time​ ​to​ ​check​ ​answers
○ If​ ​you​ ​don’t​ ​know​ ​how​ ​to​ ​answer​ ​a​ ​question,​ ​come​ ​back​ ​to​ ​it
Dealing​ ​with​ ​stress
● Practice​ ​deep​ ​breathing
● Stay​ ​positive
● Remind​ ​yourself​ ​of​ ​your​ ​past​ ​successes​ ​(but​ ​don’t​ ​get​ ​cocky)

Joanna​ ​Griffith​ ​(2017)



---