Abinda za a karanta a gaba

• Gabatarwa (sashe 2.1)

Wata irin lura da ba a haɗa a cikin wannan babi shi ne ethnography. Don ƙarin bayani game da dabi'u a cikin wurare na dijital, duba Boellstorff et al. (2012) , kuma don ƙarin bayani game da labarun da aka yi a cikin na'ura mai kwakwalwa da ta jiki, ga Lane (2016) .

• Babban bayanai (sashe na 2.2)

Babu wani maƙasudin ra'ayi ɗaya na "babban bayanai," amma ma'anoni da yawa suna neman mayar da hankali kan "3 Vs": girman, iri-iri, da kuma gudu (misali, Japec et al. (2015) ). Dubi De Mauro et al. (2015) don nazari na ma'anar.

Hakanan na hada bayanai na gwamnati a cikin jinsin bayanan babban abu ne, duk da haka wasu sunyi wannan shari'ar, ciki har da Legewie (2015) , Connelly et al. (2016) , da Einav and Levin (2014) . Don ƙarin bayani dangane da muhimmancin bayanan gwamnati don bincike, duba Card et al. (2010) , Adminstrative Data Taskforce (2012) , da Grusky, Smeeding, and Snipp (2015) .

Domin la'akari da bincike na gudanarwa daga cikin tsarin kididdiga na gwamnati, musamman Ƙungiyar Ƙididdigar Amurka, ga Jarmin and O'Hara (2016) . Domin nazari na tsawon lokaci game da bincike na bincike na Tarihi a Statistics Sweden, ga Wallgren and Wallgren (2007) .

A cikin babi, na taƙaice kwatanta binciken gargajiya irin su General Social Survey (GSS) tare da tushen bayanan yanar gizon kamar Twitter. Don kwatantaccen hankali a tsakanin nazarin gargajiya da bayanan kafofin watsa labarai, duba Schober et al. (2016) .

• Abubuwan halaye na manyan bayanai (sashe na 2.3)

Wadannan halaye 10 na manyan bayanai an bayyana su a hanyoyi daban-daban ta hanyoyi daban daban. Rubutun da ya rinjayi tunanin na game da wadannan batutuwa sun haɗa da Lazer et al. (2009) , Groves (2011) , Howison, Wiggins, and Crowston (2011) , boyd and Crawford (2012) , SJ Taylor (2013) , Mayer-Schönberger and Cukier (2013) , Golder and Macy (2014) , Ruths and Pfeffer (2014) , Tufekci (2014) , Sampson and Small (2015) , K. Lewis (2015b) , Lazer (2015) , Horton and Tambe (2015) , Japec et al. (2015) , da Goldstone and Lupyan (2016) .

A cikin wannan babi, Na yi amfani da kalmar kallo na dijital , wanda ina tsammanin yana da tsaka tsaki. Wani shahararren lokaci na dijital shine ƙafar ƙafafunni (Golder and Macy 2014) , amma kamar yadda Hal Abelson, Ken Ledeen, da Harry Lewis (2008) nuna, wani lokaci mafi dacewa shine watakila yatsan digital . Lokacin da ka kirkiro takalmin kafa, ka san abin da ke faruwa kuma ba za a iya gano hankalinka ba. Hakanan ba gaskiya ba ne saboda yanayin da kake ciki. A gaskiya ma, kuna barin alamu duk lokacin da kuke da kwarewa sosai. Kuma, ko da yake waɗannan alamu ba su da suna a kan su, za su iya sau da yawa a mayar da su a gare ku. A wasu kalmomi, sun kasance kamar yatsun hannu: wanda ba a ganuwa da kuma gano kansa.

• Babban (sashi na 2.3.1)

Don ƙarin bayani game da dalilin da yasa manyan ɗakunnan bayanai suka ba da matsala masu rikitarwa, duba M. Lin, Lucas, and Shmueli (2013) da McFarland and McFarland (2015) . Wa] annan al'amurra ya kamata jagorantar masu bincike su mayar da hankali ga muhimmancin muhimmanci maimakon mahimmanci.

Don ƙarin bayani game da yadda Raj Chetty da abokan aiki suka sami damar shiga takardun haraji, duba Mervis (2014) .

Ƙididdiga masu yawa za su iya ƙirƙirar matsalolin kwamfuta waɗanda ba su da damar haɗin kwamfuta daya. Sabili da haka, masu bincike suna yin musayar bayanai a kan manyan bayanai da yawa sukan shimfiɗa aikin a kan komfuta da dama, wani tsari wanda ake kira layi dayawa . Don gabatarwar zuwa layi daya shirin, musamman harshen da ake kira Hadoop, ga Vo and Silvia (2016) .

• Kullum-kan (sashi na 2.3.2)

Yayin da ake yin la'akari akai-akai, yana da muhimmanci muyi la'akari da ko kuna kwatanta daidai wannan mutane a lokaci ko kuma kuna kwatanta wasu kungiyoyi na mutane; ga misali, Diaz et al. (2016) .

• Ƙaƙasasshe (sashen ¿sec: ba mai amsawa ba? )

Wani littafi mai ban sha'awa a kan matakan da ba shi da tushe shine Webb et al. (1966) . Misalai a cikin wannan littafin sun kasance farkon zamani, amma har yanzu suna haskakawa. Don misalai na mutane suna canza dabi'arsu saboda kasancewa na kulawa, duba Penney (2016) da Brayne (2014) .

Rashin hankali yana da alaƙa da abin da masu bincike suke kira sakamakon binciken (Orne 1962; Zizzo 2010) da kuma Hawthorne sakamako (Adair 1984; Levitt and List 2011) .

• Ba a cika (sashi na 2.3.4)

Don ƙarin bayani a kan rikodin, duba Dunn (1946) da Fellegi and Sunter (1969) (tarihi) da Larsen and Winkler (2014) (zamani). Haka kuma an samo irin wannan matsala a cikin kimiyyar kwamfuta a karkashin sunaye kamar haɓaka bayanai, alamar misali, daidaitattun sunan, ƙwaƙwalwar haɓaka, da kuma rikodin rikodin rikodin (Elmagarmid, Ipeirotis, and Verykios 2007) . Har ila yau, akwai hanyoyin tsare-tsaren tsare-tsare don yin rikodin haɗin da ba su buƙatar watsa bayanai na mutum (Schnell 2013) . Facebook kuma ya ƙaddamar da wani tsari don danganta rubutun su don yin rikici; Anyi wannan ne don kimanta gwajin da zan fada maka a cikin babi na 4 (Bond et al. 2012; Jones et al. 2013) .

Don ƙarin bayani a kan gine-ginen gini, dubi babi na 3 na Shadish, Cook, and Campbell (2001) .

• M (sashi na 2.3.5)

Don ƙarin bayani game da lakabin bincike na AOL, duba Ohm (2010) . Ina bayar da shawara game da haɗin kai tare da kamfanoni da gwamnatoci a babi na 4 lokacin da na bayyana gwaje-gwaje. Yawancin marubuta sun bayyana damuwa game da bincike da ke dogara da bayanan da ba za a iya fahimta ba, ga Huberman (2012) da kuma boyd and Crawford (2012) .

Daya kyau hanya domin jami'a masu bincike don saya data access ne ya yi aiki a wani kamfanin a matsayin ɗalibin kwalejin likita ko ziyartar bincike. Bugu da ƙari, kunna data access, wannan tsari zai taimaka da bincike ƙarin koyo game da yadda data aka halitta, wanda yake shi ne da muhimmanci ga analysis.

Game da samun damar shiga bayanai na gwamnati, Mervis (2014) tattauna yadda Raj Chetty da abokan aiki suka sami damar yin amfani da bayanan haraji da aka yi amfani da su a cikin bincike kan zamantakewar zamantakewa.

• Ba wakili (sec: ba wakili)

Don ƙarin bayani game da tarihin "wakilci" a matsayin manufar, duba Kruskal and Mosteller (1979a) , Kruskal and Mosteller (1979b) , Kruskal and Mosteller (1979c) , kuma Kruskal and Mosteller (1980) .

Abubuwan nawa game da aikin Snow da aikin Doll da Hill sun kasance kaɗan. Don ƙarin bayani akan aikin Snow a kan kwalara, duba Freedman (1991) . Don ƙarin bayani akan Birtaniya Doctors Nazarin duba Doll et al. (2004) da kuma Keating (2014) .

Mutane da yawa masu bincike za su yi mamakin sanin cewa ko da yake Doll da Hill sun tattara bayanai daga likitoci mata da kuma daga likitoci a kasa da 35, sun yi ganganci ba su yi amfani da wannan bayanan ba a farkon bincike. Kamar yadda suke jaddada: "Tun da cutar kanjamau ta fi dacewa a cikin mata da maza a kasa da 35, ana iya samun alamun amfani a waɗannan kungiyoyi har tsawon shekaru masu zuwa. A cikin wannan rahoto na farko mun sanya hankali ga mutane masu shekaru 35 da sama. " Rothman, Gallacher, and Hatch (2013) , wanda ke da maƙalari mai taken" Me ya sa za a guji wakilci, "ya yi karin bayani game da muhimmancin Ƙirƙirar ƙirƙirar bayanai ba tare da lissafi ba.

Ƙungiyar ba ta da mahimmanci shine babban matsala ga masu bincike da gwamnatocin da suke so su yi maganganu game da dukkanin jama'a. Wannan ƙananan damuwa ne ga kamfanoni, wanda yawanci aka mayar da hankali ga masu amfani da su. Don ƙarin bayani a kan yadda Statistics Netherlands yayi la'akari da batun batun ƙaddamar da manyan bayanai, duba Buelens et al. (2014) .

Don misalai na masu bincike suna nuna damuwa game da wadanda ba wakilci ba ne na tushen manyan bayanai, gani boyd and Crawford (2012) , K. Lewis (2015b) , da Hargittai (2015) .

Don ƙarin kwatanci game da burin binciken binciken zamantakewa da bincike na annoba, gani Keiding and Louis (2016) .

Don ƙarin bayani a kan ƙoƙarin yin amfani da Twitter don yin bayani game da masu jefa ƙuri'a, musamman ma batun daga zaben Jamus na 2009, duba Jungherr (2013) da kuma Jungherr (2015) . Daga bisani zuwa aikin Tumasjan et al. (2010) masu bincike a ko'ina cikin duniya sunyi amfani da hanyoyi masu amfani da su-kamar yin amfani da maganganu don gane bambanci tsakanin maganganu masu kyau da kuma mummunan ra'ayi na jam'iyyun - don inganta yawan bayanan Twitter don hango nesa da irin wadannan zabuka (Gayo-Avello 2013; Jungherr 2015, chap. 7.) . Ga yadda Huberty (2015) taƙaita sakamakon wadannan ƙoƙari na hango hasashen zaben:

"Duk hanyoyin da aka sani da aka tsara a kan kafofin watsa labarun sun gaza lokacin da ake biyan bukatun kaddamar da zabe na gaba. Wadannan kasawa sun kasance ne saboda ainihin kaddarorin kafofin watsa labarun, maimakon ƙwarewa ko matsala masu algorithmic. A takaice dai, kafofin watsa labarun ba su da, kuma ba za su taba ba, suna ba da kwanciyar hankali, ba tare da nuna bambanci ba, hoto na wakiltar za ~ en; kuma saukaka samfurori na kafofin watsa labarun ba su da cikakkun bayanai don magance wadannan matsaloli ba a baya. "

A cikin babi na 3, zan bayyana samfurin da lissafi a cikin cikakken bayani. Ko da koda bayanan bayanai ba su da wata ma'ana, a wasu yanayi, za su iya zama masu la'akari don samar da kimantaccen kimantawa.

• Drifting (sashe ¿sec: drifting? )

Drift tsarin yana da wuyar ganin daga waje. Duk da haka, aikin na MovieLens (tattaunawa akan su a cikin sura na 4) an gudanar da shi har tsawon shekaru 15 da wani rukunin bincike na ilimi. Ta haka ne, sun sami damar rubutawa da kuma rarraba bayanai game da yadda tsarin ya samo asali a tsawon lokaci kuma yadda hakan zai iya tasiri tasirin (Harper and Konstan 2015) .

Wasu malaman sun mayar da hankalinsu a kan Twitter: Liu, Kliman-Silver, and Mislove (2014) da kuma Tufekci (2014) .

Ɗaya daga cikin hanyoyin da za a magance yawancin jama'a shi ne ƙirƙirar panel na masu amfani, wanda ya ba da damar masu bincike suyi nazarin waɗannan mutane a tsawon lokaci, ga Diaz et al. (2016) .

• Algorithmically kunya (sashe ¿sec: algorthmically-kunya? )

Na fara jin kalmar "algorithmically confounded" da Jon Kleinberg yayi a cikin magana, amma rashin alheri ban tuna lokacin ko inda aka ba da labarin ba. A karo na farko da na ga lokacin a buga shi ne a Anderson et al. (2015) , wanda ke da mahimmanci tattaunawa game da yadda algorithms da aka yi amfani da su ta hanyar shafukan yanar gizo na iya haifar da damar masu bincike na iya amfani da bayanai daga waɗannan shafukan yanar gizo don nazarin abubuwan da ake so. Wannan damuwa ya taso ne daga K. Lewis (2015a) a cikin amsa ga Anderson et al. (2014) .

Bugu da ƙari, Facebook, Twitter kuma ya ba da shawarar mutane don masu amfani su bi bisa ga ra'ayin ƙaddamar ƙaddamarwa; ga Su, Sharma, and Goel (2016) . Saboda haka matakin ƙulli na triadic a cikin Twitter shi ne haɗuwa da wasu dabi'un ɗan adam zuwa ƙulli na taya da kuma wasu halayen algorithmic don bunkasa ƙulli na taya.

Don ƙarin bayani game da aiki-musamman ma ra'ayin cewa wasu masana kimiyya na zamantakewar al'umma "na'urori ba kyamarori ba" (watau, suna siffar duniya maimakon bayyana shi) - Mackenzie (2008) .

• Dirty (sashi na 2.3.9)

Hukumomi na gwamnati sun kira bayanan tsaftacewa na gyara bayanai . De Waal, Puts, and Daas (2014) bayyana fasali na bayanan kididdigar da suka samo don nazarin binciken da kuma nazarin yadda za su dace da asali na manyan bayanai, da kuma Puts, Daas, and Waal (2015) gabatar da wasu ra'ayoyin don karin masu sauraro.

Don ƙarin bayani game da burbushin zamantakewa, duba Ferrara et al. (2016) . Ga wasu misalai na nazarin da aka mayar da hankali a kan gano spam a Twitter, duba Clark et al. (2016) da Chu et al. (2012) . A ƙarshe, Subrahmanian et al. (2016) bayyana sakamakon DARPA Twitter Bot Challenge, wani haɗin gwiwar da aka ƙera don kwatanta hanyoyin da za a gano buƙatun a kan Twitter.

• Mai hankali (sashi na 2.3.10)

Ohm (2015) bincika binciken da suka gabata a kan ra'ayin mahimman bayanai kuma ya ba da gwaji mai yawa. Hanyoyi guda hudu da ya gabatar yana da mummunar cutar, yiwuwar cutar, kasancewar dangantaka ta sirri, da kuma yadda haɗarin ya nuna damuwa da damuwa.

• Ƙidaya abubuwa (sashe 2.4.1)

Nazarin takardun haraji na Farber a New York ya dogara akan binciken farko da Camerer et al. (1997) wanda yayi amfani da samfurori daban-daban daban daban na takarda takarda. Wannan binciken da aka yi a baya ya gano cewa direbobi sun kasance masu daukar nauyin da suke da manufa: sun yi aiki a ƙasa a kwanakin da farashin su ya fi girma.

A cikin aiki na gaba, Sarki da abokan aiki sun sake binciko bincike kan layi a kasar Sin (King, Pan, and Roberts 2014, [@king_how_2016] ) . Domin wata hanyar da ta dace game da yin la'akari da bincike kan layi a kasar Sin, duba Bamman, O'Connor, and Smith (2012) . Don ƙarin bayani a kan hanyoyin da aka yi amfani da su a cikin King, Pan, and Roberts (2013) don kwatanta jinin kujerun miliyoyin 11, duba Hopkins and King (2010) . Don ƙarin bayani game da ilmantarwa, duba James et al. (2013) (ƙananan fasahar) da Hastie, Tibshirani, and Friedman (2009) (mafi fasaha).

• Bayani da kuma bayyanawa yanzu (sashe 2.4.2)

Bayani mai girma shine babban ɓangaren kimiyya na masana'antu (Mayer-Schönberger and Cukier 2013; Provost and Fawcett 2013) . Ɗaya daga cikin sharuddan da ake amfani da shi ta hanyar masu bincike na zamantakewar al'umma shine fadin alƙaluma; ga, misali, Raftery et al. (2012) .

Google Flu Trends ba shine aikin farko ba don amfani da bayanan bincike don yanzu yaduwar cutar. A gaskiya ma, masu bincike a Amurka (Polgreen et al. 2008; Ginsberg et al. 2009) da kuma Sweden (Hulth, Rydevik, and Linde 2009) sun gano cewa wasu sharuɗan bincike (misali, "mura") sun annabta lafiyar lafiyar jama'a bayanai kafin a sake shi. Bayan haka, wasu ayyukan da yawa sunyi ƙoƙari don amfani da bayanai na layi don gano magunguna; duba Althouse et al. (2015) don bita.

Bugu da ƙari, yin amfani da bayanan layi don gano hangen nesa na kiwon lafiya, an samu babban aiki tare da amfani da bayanan Twitter don tantance sakamakon zaben; don dubawa Gayo-Avello (2011) , Gayo-Avello (2013) , Jungherr (2015) (babi na 7), da Huberty (2015) . Kasancewa da alamun tattalin arziki, irin su samfur na gida (GDP), ma a cikin bankunan tsakiya, dubi Bańbura et al. (2013) . tebur 2.8 ya haɗa da wasu misalai na nazarin da suka yi amfani da wasu nau'ikan tallace-tallace don ganin hangen nesa a duniya.

A karshe, Jon Kleinberg da abokan aiki (2015) sun nuna cewa matsalolin da ke faruwa a cikin rikice-rikice sun fada cikin nau'o'i biyu, masu rarrabuwa daban-daban kuma cewa masana kimiyyar zamantakewa sun kula da daya kuma sun watsar da juna. Ka yi la'akari da daya daga cikin manufofi, zan kira ta Anna, wanda ke fuskantar fari kuma dole ne ya yanke shawara ko ya hayar da shaman don yin rawar ruwa don kara yawan ruwan sama. Wani mai tsara manufar, zan kira shi Betty, dole ne in yanke shawarar yin laima don yin aiki don kaucewa yin rigakafi kan hanya. Duk Anna da Betty na iya yin shawara mafi kyau idan sun fahimci yanayin, amma suna bukatar su san abubuwa daban-daban. Anna ya kamata ya fahimci ko raye-raye na ruwa ya ja ruwa. Betty, a gefe guda, ba ya bukatar fahimtar wani abu game da causality; Tana buƙatar cikakken bayani. Masu bincike na zamantakewa sau da yawa sukan mayar da hankali ga matsalolin kamar Anna wanda Kleinberg da abokan aiki suke kira "damun raye-raye" matsalolin siyasa-domin sun ƙunshi tambayoyi game da causality. Tambayoyi kamar wanda Betty-abin da Kleinberg da abokan aiki suka fuskanta suna kira "matsaloli-kamar" matsaloli na siyasa-na iya zama mahimmanci ma, amma sun sami kulawa da yawa daga masu bincike na zamantakewa.

• Gwaran gwaje-gwaje (sashi 2.4.3)

Labaran ta PS na Siyasa Siyasa na da taron tattaunawa game da manyan bayanai, ƙaddamar da damuwa, da ka'idar ka'ida, da kuma Clark and Golder (2015) taƙaita kowace gudunmawa. Jaridar The Proceedings of the National Academy of Sciences of the United States of America na da taron tattaunawa game da ƙaddamar da ƙwarewa da kuma manyan bayanai, kuma Shiffrin (2016) taƙaita kowace gudummawa. Don hanyoyin dabarun gwaji game da ƙoƙari don bincika gwaji na ainihi a cikin manyan bayanai, duba Jensen et al. (2008) , Sharma, Hofman, and Watts (2015) , da Sharma, Hofman, and Watts (2016) .

Dangane da gwaje-gwaje na halitta, Dunning (2012) samar da gabatarwa, yin nazari na tsawon lokaci tare da misalai. Domin ra'ayin da ya dace game da gwaje-gwaje na halitta, duba Rosenzweig and Wolpin (2000) (tattalin arziki) ko Sekhon and Titiunik (2012) (kimiyyar siyasa). Deaton (2010) da Heckman and Urzúa (2010) suna jayayya cewa mayar da hankalin gwaje-gwaje na al'ada zai iya haifar da masu bincike don mayar da hankali ga ƙaddamar da tasiri masu tasiri; Imbens (2010) la'akari da waɗannan hujjoji tare da ra'ayi mafi kyau game da muhimmancin gwaji na halitta.

Lokacin da ya kwatanta yadda mai bincike zai iya tafiya daga kimantawa da sakamakon yin aiki a kan aikin hidima, na kwatanta wata fasaha da ake kira ƙananan kayan aiki . Imbens and Rubin (2015) , a cikin surori 23 da 24, samar da gabatarwar da amfani da irin caca misali. Sakamakon aikin soja a kan compliers wani lokaci ana kiran shi da matsakaicin sakamako na lalacewa (CAcE) da kuma wani lokacin magungunan magani na gida (LATE). Sovey and Green (2011) , Angrist and Krueger (2001) , da kuma Bollen (2012) suna ba da damar yin amfani da kayan aiki a cikin kimiyyar siyasa, tattalin arziki, da zamantakewar zamantakewa, da kuma Sovey and Green (2011) ba da Sovey and Green (2011) "masu karatu" kimantawa da karatu ta amfani da maɓalli masu aiki.

Ya bayyana cewa wasan kwaikwayo na shekarar 1970 ba shine, a gaskiya ma baƙi ba; akwai ƙananan ƙetare daga tsabta baƙi (Fienberg 1971) . Berinsky and Chatfield (2015) yi jayayya cewa wannan ƙananan ƙetare ba muhimmiyar mahimmanci ba ne kuma ya tattauna muhimmancin yin gyaran ƙira.

Dangane da daidaitawa, duba Stuart (2010) don nazari mai kyau, da kuma Sekhon (2009) don sake dubawa. Don ƙarin bayani game da daidaituwa kamar nau'i-nau'i, duba Ho et al. (2007) . Samun cikakkiyar wasa ga kowane mutum yana da wuya sau da yawa, kuma wannan yana gabatar da wasu abubuwa masu ban mamaki. Na farko, idan matakan da ba su samuwa ba, masu bincike suna bukatar yanke shawarar yadda za su auna nisa tsakanin raka'a biyu kuma idan nesa da aka isa ya isa. Matsalar ta biyu ta taso idan masu bincike suna so su yi amfani da matakai masu yawa don kowane hali a cikin ƙungiyar kulawa, saboda wannan zai haifar da kimantaccen ƙayyade. Duk waɗannan batutuwa, da sauransu, an kwatanta dalla-dalla a babi na 18 na Imbens and Rubin (2015) . Duba kuma Sashe na II na ( ??? ) .

Dubi Dehejia and Wahba (1999) don misali inda hanyoyin daidaitawa sun iya samar da kimantawa kamar wadanda daga gwajin gwajin da aka samu. Amma, duba Arceneaux, Gerber, and Green (2006) da Arceneaux, Gerber, and Green (2010) don misalai inda hanyoyin daidaitawa sun kasa samar da alamar gwaji.

Rosenbaum (2015) da Hernán and Robins (2016) ba da shawarwari don gano kwarewan amfani a cikin manyan bayanai.