182
edits
m
→Discoveries
mNo edit summary |
m (→Discoveries) |
||
| Line 10: | Line 10: | ||
=== Test Sample === | === Test Sample === | ||
[[File:Lasswade Parish Baptisms Index Accuracy - Becky Adamson.png|thumb|537x537px]] | |||
A test sample of records were extracted from the register of baptisms for Lasswade Parish. A five-year period from 1750-1754 was selected from which all entries were extracted. This resulted in 371 records that were searched for in the three indexes: SCRI, SB&B, and SP. SCRI was missing two entries - a 99% accuracy. The SB&B fell short at 68%, only including 253 of the 371 entries. The SP index was 99.7% accurate, missing only one entry. The preliminary study suggested the number of missing records in the SB&B index would be significant, creating a need for the SCRI. It also suggested data trends which could potentially define the parameters surrounding which entries are missing in the SB&B. | A test sample of records were extracted from the register of baptisms for Lasswade Parish. A five-year period from 1750-1754 was selected from which all entries were extracted. This resulted in 371 records that were searched for in the three indexes: SCRI, SB&B, and SP. SCRI was missing two entries - a 99% accuracy. The SB&B fell short at 68%, only including 253 of the 371 entries. The SP index was 99.7% accurate, missing only one entry. The preliminary study suggested the number of missing records in the SB&B index would be significant, creating a need for the SCRI. It also suggested data trends which could potentially define the parameters surrounding which entries are missing in the SB&B. | ||
=== Data Extracted for Study === | === Data Extracted for Study === | ||
| Line 31: | Line 38: | ||
A comparison between the FamilySearch indexes and SCRI showed that SCRI was more complete than the online indexes at ''FamilySearch''. SCRI's 98.9% accuracy bested the 85.6% accuracy found in the SB&B for baptisms and the 83% accuracy for marriage data found in the SM. When comparing SCRI's data to the indexes at ''Scotlandspeople'', they were essentially equals. Data from SP showed their indexes to be 99.2% complete in baptismal records and 97.1% in marriage data. A margin of error was not calculated into the analysis of this study. | A comparison between the FamilySearch indexes and SCRI showed that SCRI was more complete than the online indexes at ''FamilySearch''. SCRI's 98.9% accuracy bested the 85.6% accuracy found in the SB&B for baptisms and the 83% accuracy for marriage data found in the SM. When comparing SCRI's data to the indexes at ''Scotlandspeople'', they were essentially equals. Data from SP showed their indexes to be 99.2% complete in baptismal records and 97.1% in marriage data. A margin of error was not calculated into the analysis of this study. | ||
[[File:Index Analysis Comparison.png|none|thumb|797x797px]] | |||
=== Discoveries === | === Discoveries === | ||
Data analysis revealed some interesting trends. From the information gathered in the test sample, it was found that in every instance that an individual was found in the FamilySearch family tree without a surname, that record was omitted from the SB&B database. From the 371 baptismal records extracted from Lasswade parish, 118 were not in SB&B. Of these, 53 records fell into this category of no surname. This data was evident in the test sample only, and not confirmed in the overall study. | [[File:Surname Example.png|thumb|660x660px|''A search for Margaret Watson resulted in this example of an extracted record entered into the Family Tree without a surname. Margaret's surname is known only because she is attached to her father.'' ]] | ||
Data analysis revealed some interesting trends. From the information gathered in the test sample, it was found that in every instance that an individual was found in the FamilySearch family tree without a surname, that record was omitted from the SB&B database. From the 371 baptismal records extracted from Lasswade parish, 118 were not in SB&B. Of these, 53 records fell into this category of no surname. This data was evident in the test sample only, and not confirmed in the overall study. | |||
Another common marker noted from information collected from both the test sample and the study showed that 54% of the missing entries in the SB&B were individuals that had church ordinance work done in the 1970s. | [[File:Missing Records by Ordinance.png|thumb|659x659px|''Because the church ordinance dates covered such a wide range, to have 54% of the records missing in one decade is significant.'' ]] | ||
Another common marker noted from information collected from both the test sample and the study showed that 54% of the missing entries in the SB&B were individuals that had church ordinance work done in the 1970s. | |||
=== Conclusion === | === Conclusion === | ||