Monday, 24 September 2012

Twitter chat transcript - quantitative methods

Thanks to everyone who joined us on Twitter to discuss quantitative methods in social media research - it was great to have a range of people involved. If you weren't able to make it, or if you want to take another look at what was said, the transcript is below - scroll to the bottom and work your way back up..

We'll be holding another Twitter chat in October and we'd love to hear your thoughts on what it should cover. Do let us know on Methodspace.



BrindlePatrick
#NSMNSS Whither qualitative research in the world of big data? A q for next time perhaps?
EmUprichard
@jess1ecat @NSMNSS Yes, I don't think we've got our heads round images yet, not in the way the digital world needs us to... #NSMNSS
EmUprichard
@NSMNSS Was v good. Think you've found a new interactive survey analysis tool - was much more fun than filling in a questionnaire! #NSMNSS
DrLukeSloan
@NSMNSS Some great ideas here - the next big challenge for Social Science and a chance to revolutionise research #NSMNSS
huwcdavies
@jess1ecat glad to find like minds - will here next time for sure #NSMNSS
jess1ecat
@huwcdavies I share that sentiment, see you next time I hope. #NSMNSS
huwcdavies
#NSMNSS I'm a Web Scientist/Sociologist from Southampton Uni - I'm a big data sceptic atm but would like to be convinced
PostPhDoc
@NSMNSS Great chat - the value of text and engagement as opposed to pure 'big data'.  #NSMNSS #sociology
NSMNSS
Bye all - see you in October's chat. And look out for us livetweeting from tomorrow's event #NSMNSS
jess1ecat
@GerryNicolaas our pleasure :) #NSMNSS
merry30
Ditto. MT @jess1ecat A5. ... how social media research is affecting our relationships with research participants. #NSMNSS
GerryNicolaas
Bye & thanks for interesting chat. Didn't think that I would take to it but pleasantly surprised! And I have new reading material... #NSMNSS
NSMNSS
Thanks for joining us. Great chat. We'll post the transcript here shortly: http://t.co/VLQZLhdQ #NSMNSS
jess1ecat
Thanks everyone for taking part, it's good to talk! We'll be blogging from tomorrow's event & holding a follow-up webinar in Oct. #NSMNSS
VickiMcDermott
RT @NSMNSS: Join us at #NSMNSS at for a lively discussion about quantitative methods & social media research– newcomers welcome 4-5p ...
GerryNicolaas
@DrLukeSloan I like your optimism and to some extent I share it. Difficulty=user profile of different soc media outlets changes. #NSMNSS
NSMNSS
Thanks everyone. Please take a moment to reintroduce yourself and leave a parting comment. #NSMNSS
merry30
@DrLukeSloan @jess1ecat @GerryNicolaas The never ending conundrum! #NSMNSS
jess1ecat
@NSMNSS @EmUprichard good one, we see a lot of focus on text and numbers but very little on the wealth of socialmedia images. #NSMNSS
EmUprichard
@NSMNSS Oops, sorry! #NSMNSS
NSMNSS
A5. #NSMNSS RT @EmUprichard: @NSMNSS How to deal with millions of digitised photos/images?
iamandyscott
A5. #NSMNSS RT @EmUprichard: @NSMNSS Time and temporality.
NSMNSS
A5. #NSMNSS RT @EmUprichard: @NSMNSS Time and temporality.
huwcdavies
@jess1ecat http://t.co/KwEH7FlM there's a lot of big data research here that requires a healthy dose of social science scepticism #NSMNSS
109deb
@huwcdavies I wonder how similar data access issues are for surveys and social media? #nsmnss
NSMNSS
@EmUprichard Remember to use the hashtag #NSMNSS or our other tweeters won't be able to see your contribution
DrLukeSloan
@GerryNicolaas Sounds like a starting point! Similar discussions happened in the 1960s regarding survey methodology - they solved it #NSMNSS
merry30
@jess1ecat @GerryNicolaas  Don't think we can really redefine "robust" altogether, but maybe in context of soc med research #NSMNSS
jess1ecat
A5. I'd be interested in discussing how social media research is affecting our relationships with research participants. #NSMNSS
PostPhDoc
@NSMNSS @EmUprichard That's a good answer for me! To go further we would have to probe & do a bit of qualitative research/analysis #NSMNSS
jess1ecat
@DrLukeSloan Ha ha, not sure I like the idea that we had to drag you in! #NSMNSS
jess1ecat
RT @NSMNSS: Short and to the point on the value of data viz #NSMNSS RT @EmUprichard: @NSMNSS No #NSMNSS
NSMNSS
Q5. Almost time to wrap up but we'd like your suggestions for topics for October's chat first. #NSMNSS
huwcdavies
A.1 The more interesting/precious the data the more likely it will be kept in a silo protected by IP e.g. Google, Tesco etc #NSMNSS
DrLukeSloan
@jess1ecat Hello! Couldn't help but get dragged in and distracted #NSMNSS
NSMNSS
Short and to the point on the value of data viz #NSMNSS RT @EmUprichard: @NSMNSS No
GerryNicolaas
@DrLukeSloan That sounds like a challenge! ;-) Random prob sampling plus additional efforts to ensure a balanced sample... #NSMNSS
jess1ecat
The heart of the issue being what meaning we can glean from analysis like that. #NSMNSS
jess1ecat
@huwcdavies interesting questions abt whether a RT means you've influenced someone, gets to the heart of the issue. #NSMNSS
DrLukeSloan
@jess1ecat @merry30 @GerryNicolaas but we could say that survey participants are only representative of survey participants #NSMNSS
jess1ecat
Hi @DrLukeSloan good to see you. #NSMNSS
huwcdavies
A2. I see many claims based on big data network graphs eg w influences x,y, & z  'cause w retweets them- we need quan to tell us why #NSMNSS
PostPhDoc
@GerryNicolaas  I think it certainly has its place - but in conjunction with qualitative methods esp. in analysing soc media data #NSMNSS
109deb
@GerryNicolaas point taken about usefulness of theory. But its a two way street. Do we have theories that will help analyse sm data? #nsmnss
DrLukeSloan
@NSMNSS The problem is deriving the necessary demographic Info to ensure that your sample is representative, but it's not impossible #NSMNSS
jess1ecat
A4. Without doubt it's adding impact, getting stats to a wider audience, but is it adding methodologically?  #NSMNSS
jess1ecat
RT @NSMNSS: Q4. Do we think that data visualisation is adding anything to existing methods of statistical analysis? #NSMNSS
NSMNSS
Sorry, that was Q4.. RT @NSMNSS: Do we think that data visualisation is adding anything to existing methods of statistical analysis? #NSMNSS
jess1ecat
@merry30 @GerryNicolaas A3. being devil's advocate here is that really true or do we need to redefine what robust means? #NSMNSS
NSMNSS
Do we think that data visualisation is adding anything to existing methods of statistical analysis? #NSMNSS
PostPhDoc
@NSMNSS Q3. as in Q1.validity-authentication of data.Depends on if representativeness is sought, which would b difficult 2 ascertain #NSMNSS
GerryNicolaas
@109deb Risk=search and ye will find. The datasets are huge. Theory driven research is vital. #NSMNSS
jess1ecat
A3. Also @LoveStats has some interesting thoughts on this see http://t.co/U84kg17G & her blog http://t.co/1T7m83M8.
 #NSMNSS
109deb
A3. Sample is of comments not people. Comments will be as representative as your search terms allowed. So skill is searching. #nsmnss
NSMNSS
RT @jess1ecat: A3. Theres an interesting blog on sampling socmed data here: http://t.co/ZsFWHVHQ shows how sampling shapes outputs. #NSMNSS
merry30
@GerryNicolaas A3. We can't! #nsmnss
iamandyscott
@gerrynicolaas A3 In all of this its important to understand the limitations of SM res. Not over-claiming will give it more strength #NSMNSS
jess1ecat
A3. There's an interesting blog on sampling socmed data here: http://t.co/Kh74RFw2 shows how sampling shapes outputs. #NSMNSS
GerryNicolaas
A3. As mentioned before - soc media users are only representative of soc media users. How do we make inferences to other pop groups? #NSMNSS
jess1ecat
@KateT_health Hi Kate do join in we're interested in everyone's thoughts. #NSMNSS
GerryNicolaas
A2 And we need agreed standards on the methods being used to source and analyse sm data and big data more generally. #NSMNSS
NSMNSS
Q3. What challenges does social media research pose for quant researchers wanting to use robust samples? #NSMNSS
jess1ecat
@109deb A2. Yes, it's increasingly popular but I wonder if the cart (great data viz) has come before the horse (useful analysis)... #NSMNSS
iamandyscott
Another big issue with big data is who owns it and getting access to it. Need to bring owners into research 'tent' #NSMNSS
KateT_health
Following the #NSMNSS thread- discussion on quantitative methods + social media data- interesting
GerryNicolaas
A2 As well as methods, we also need theory. #NSMNSS
jess1ecat
A2. Big data can be used for public good, of that I'm sure (data journalism e.g.) but we still need to crack issues of quality. #NSMNSS
109deb
A2 - Not sure sm currently being used to answer kinds of Qs that surveys would answer. Or that commissioners know what to do with it #nsmnss
GerryNicolaas
RT @iamandyscott: @GerryNicolaas @jess1ecat Need to understand structure before we can analyse it. Difficult when always changing. #NSMNSS
PostPhDoc
@NSMNSS Q2. raises inevitable questions of validity & authenticity. Disclaimers on difficulties in verification become commonplace #NSMNSS
iamandyscott
@GerryNicolaas @jess1ecat Chicken & egg. Need to understand the structure before we can analyse it. Difficult when always changing. #NSMNSS
GerryNicolaas
@jess1ecat Indeed! Availability of big data precedes availability of suitable quant methods for analysing complex data structure. #NSMNSS
jess1ecat
RT @GerryNicolaas: A2 Risk=analyse big data because its available rather than it being suitable to answer research question. >Agreed #NSMNSS
jess1ecat
A2. The volumes of data are huge and our control over their shape is limited. I don't think bigger necess means more useful.  #NSMNSS
GerryNicolaas
A2 Risk=analyse big data because it's available rather than it being suitable to answer research question. #NSMNSS
109deb
@109deb A1. I think we need qual and quant methods, depending on what the analysis questions are.  #nsmnss
NSMNSS
Q2. Big data - is it changing the way we do research. If so, how? #NSMNSS
NSMNSS
Interesting stuff. While we're pondering that, let me put the next question to you, which might help to add some context #NSMNSS
jess1ecat
@VickiMcDermott hi Vicki nice to see you. #NSMNSS
GerryNicolaas
@PostPhDoc Hi! Would you therefore say that quant methods shouldn't be uesd to analyse soc media data? #NSMNSS
jess1ecat
A1. There's has been an explosion in the use of quant data from digital streams what is it telling us that nondigital quant doesn't? #NSMNSS
jess1ecat
MT: @PostPhDoc: Overall quality of data-social media is an expressive medium. Quant data alone wont provide the complete narrative #NSMNSS
GerryNicolaas
@109deb A1 If so, the use of soc media data is quite restricted. My impression is that researchers want to use the data more widely. #NSMNSS
PostPhDoc
Q1. Overall quality of data-social media is an expressive medium. Quant data alone won't suffice in providing the complete narrative #NSMNSS
G_Morrell
RT @NSMNSS: Q1. What are the biggest methodological challenges when working with quantitative social media data? #NSMNSS
G_Morrell
Hi, Gareth Morrell, involved in the running the network. Others, please do join us!
 #NSMNSS
VickiMcDermott
RT @NSMNSS: Q1. What are the biggest methodological challenges when working with quantitative social media data? #NSMNSS
109deb
@GerryNicolaas we are measuring the behaviour of using social media and either providing or passing on information #nsmnss
GerryNicolaas
A1 I agree with Andy - soc media users are not representative of any larger group. They are only representative of soc media users.
 #NSMNSS
iamandyscott
A1. Huge problem with sampling - lack of demographic info on most people's profiles makes this even harder. #NSMNSS
GerryNicolaas
A1 Compared to surveys, soc media data are conversations rather than responses to standardised questions. What are we measuring? #NSMNSS
NSMNSS
Q1. What are the biggest methodological challenges when working with quantitative social media data? #NSMNSS
NSMNSS
Great to see you all here. Let's get started. #NSMNSS
GerryNicolaas
RT @jess1ecat: We're a little NatCen heavy today, do join us if youre interested in quant methods for social media research #NSMNSS #NSMNSS
109deb
My interest is in the measurement of behaviour and whether social media offers 'better' measurement tools than surveys. #nsmnss
jess1ecat
Looks like we're a little NatCen heavy today, do join us if you're interested in quant methods for social media research #NSMNSS #NSMNSS
iamandyscott
Hi, Andy Scott - Researcher at @NatCen #NSMNSS
109deb
Hi Debbie Collins, Head of Questionnaire Development and Testing at NatCen Social Research. #NSMNSS #nsmnss
NSMNSS
Remember to use #NSMNSS in your reply so everyone can see it and add A1, A2 etc as you address each question #NSMNSS
NSMNSS
There's a summary of what we'll be talking about today over on our blog, btw http://t.co/Q4283CnC #NSMNSS
jess1ecat
*waves* hi Gerry good to see you here. #NSMNSS
GerryNicolaas
I forgot to mention that I'm from NatCen Social Research. #NSMNSS
jess1ecat
RT @NSMNSS: Welcome along everyone! Please introduce yourselves, tell us where youre from and what you interest in the topic is. #NSMNSS
GerryNicolaas
Hi. I'm a survey methodologist who disagrees that soc media data will replace surveys. Supplement - probably. I'm here to find out. #NSMNSS
NSMNSS
Welcome along everyone! Please introduce yourselves, tell us where you're from and what you interest in the topic is. #nsmnss
jess1ecat
Hi I'm Kandy, @NatCen co-network lead. Great to see you all here. #NSMNSS
NSMNSS
Our next tweetchat is starting in 5 minutes, follow #NSMNSS to join the debate - don't be late! #nsmnss

No comments:

Post a Comment